{ "cells": [ { "cell_type": "markdown", "id": "7fc9a3a8", "metadata": {}, "source": [ "\n", "```{math}\n", "\n", "\\renewcommand{\\Pr}{\\mathrm{Pr}}\n", "\\renewcommand{\\positive}{\\checkmark}\n", "\\newcommand{\\Binomial}{\\mathrm{Binomial}}\n", "\\newcommand{\\Beta}{\\mathrm{Beta}}\n", "\\renewcommand{\\B}{\\mathrm{B}}\n", "\\newcommand{\\Pynchon}{\\mathrm{Pynchon}}\n", "\\newcommand{\\distro}[1]{\\text #1}\n", "\\newcommand{\\E}[1]{\\mathbb{E}{\\text #1}}\n", "\\newcommand{\\Var}[1]{\\mathbb{Var}{\\text #1}}\n", "\\newcommand{\\and}{\\ \\mathrm{and}\\ }\n", "```" ] }, { "cell_type": "code", "execution_count": 1, "id": "9f749b48", "metadata": { "tags": [ "remove-cell" ] }, "outputs": [], "source": [ "import matplotlib.pyplot as plt\n", "plt.style.use(\"../styles/hda.mplstyle\")" ] }, { "cell_type": "markdown", "id": "d9aa3f04", "metadata": {}, "source": [ "(chp-intro-probability)=\n", "# Introduction to Probability\n", "\n", "\n", "Probabilities are used in everyday speech to describe uncertain events and\n", "degrees of belief about past and future states of affairs. A newspaper writer\n", "might, for example, write about the likely outcome of an election held the\n", "following day, or a student might talk about the their chance of having received\n", "a passing score on a test they just completed. This chapter discusses the use of\n", "probabilities in the context of one account of what it means to learn from\n", "experience: Bayesian inference. Bayesian inference provides both (1) an\n", "interpretation of probabilities, and (2) a prescription for how to update\n", "probabilities in light of evidence (i.e., how to learn from experience). Both\n", "these items are important. Treating probability statements as statements about\n", "degree of belief is useful whenever we need to communicate or compare our\n", "judgment of the credibility of a hypothesis (or model) with the judgments of others,\n", "something that happens all the time in the course of historical research. And\n", "having a method of learning from evidence is important because it is useful to be able\n", "to assess how evidence, in particular new evidence, supports or undermines theories.\n", "\n", "The goal of this chapter is to illustrate Bayesian inference as applied to the\n", "task of authorship attribution. The chapter begins with an informal discussion\n", "of probabilities and Bayesian inference. A fictional case of authorship\n", "attribution is used to prepare readers for a full treatment later in the\n", "chapter. This introduction is followed by a formal characterization of\n", "probabilities and Bayes's rule. The introduction should leave readers familiar\n", "with the basic vocabulary encountered in data analysis in the humanities and\n", "allied social sciences. This formal introduction is followed by a return to the\n", "task of authorship attribution. The bulk of the chapter is devoted to a detailed\n", "case study examining the classic case of disputed authorship of several essays in *The Federalist Papers*.\n", "\n", "One note of warning: this chapter does not offer a complete introduction to probability. An appreciation of the material in this chapter requires some\n", "familiarity with standard discrete probability distributions. This kind of familiarity is typically obtained in an introductory mathematics or statistics course devoted to probability.\n", "We discuss appropriate preliminary readings at the end of this chapter.\n", "For many kinds of humanities data, one specific discrete probability distribution---one featured in this chapter---is of great importance: the negative binomial distribution. Indeed,\n", "facility with the negative binomial distribution counts as an essential skill\n", "for anyone interested in modeling word frequencies in text documents.\n", "The negative binomial distribution is the simplest probability distribution which provides a plausible model of word frequencies observed in English-language documents {cite:p}`church1995poisson`.\n", "The only kinds of words which can be credibly modeled with more familiar distributions (e.g., normal, binomial, or Poisson) are those which are extremely frequent or vanishingly rare.\n", "\n", "(sec-intro-probability-introduction)=\n", "## Uncertainty and Thomas Pynchon\n", "\n", "To motivate the Bayesian task of learning from evidence, consider the following\n", "scenario. It has been suggested that [Thomas\n", "Pynchon](https://en.wikipedia.org/wiki/Thomas_Pynchon), a well-known American\n", "novelist, may have written under a pseudonym during his long career (e.g.,\n", "{cite:t}`winslow2015fiction`). Suppose the probability of a work of literary fiction\n", "published between 1960 and 2010 being written by Thomas Pynchon (under his own\n", "name or under a pseudonym) is 0.001 percent, i.e., 1 in 100,000. Suppose,\n", "moreover, that a stylometric test exists which is able to identify a novel as\n", "being written by Pynchon 90 percent of the time (i.e., the true positive\n", "rate---\"sensitivity\" in some fields---equals 0.9). One percent of the time, however, the test mistakenly\n", "attributes the work to Pynchon (i.e., the false positive rate equals\n", "0.01). In this scenario, we assume the test works as described; we might imagine\n", "Pynchon himself vouches for the accuracy of the test or that the test invariably\n", "exhibits these properties over countless applications. Suppose a novel (written\n", "by someone other than Pynchon) published in 2010 tests positive on the\n", "stylometric test of Pynchon authorship. What is the probability that the novel\n", "was penned by Pynchon?\n", "\n", "One answer to this question is provided by Bayes's rule, which is given below and\n", "whose justification will be addressed shortly:\n", "\n", "\n", "\n", "\\begin{equation}\\label{eq:pynchon-bayes}\n", "\\Pr(\\Pynchon|\\positive) = \\frac{\\Pr(\\positive|\\Pynchon)\\Pr(\\Pynchon)}{\\Pr(\\positive|\\Pynchon)\\;\\Pr(\\Pynchon) + \\Pr(\\positive|\\neg \\Pynchon)\\;( 1 - \\Pr(\\Pynchon) )}\n", "\\end{equation}\n", "\n", "where $\\positive$ indicates the event of the novel testing positive and $\\Pynchon$ indicates the event of the novel having been written by Pynchon.\n", "The preceding paragraph provides us with values for all the quantities on the\n", "right-hand side of the equation, where we have used the expression $\\Pr(A)$ to indicate\n", "the probability of the event $A$ occurring." ] }, { "cell_type": "code", "execution_count": 2, "id": "348fa603", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "0.000899199712256092\n" ] } ], "source": [ "pr_pynchon = 0.00001\n", "pr_positive = 0.90\n", "pr_false_positive = 0.01\n", "print(pr_positive * pr_pynchon / (pr_positive * pr_pynchon +\n", " pr_false_positive * (1 - pr_pynchon)))" ] }, { "cell_type": "markdown", "id": "8eb95e02", "metadata": {}, "source": [ "Bayes' rule produces the following answer: the probability that Pynchon is\n", "indeed the author of the novel given a positive test is roughly one tenth of\n", "one percent. While this is considerably higher than the baseline of one hundredth\n", "of one percent, it seems underwhelming given the stated accuracy (90 percent)\n", "of the test.\n", "\n", "```{note}\n", " A less opaque rendering of this rule relies on the use of natural\n", " frequencies. Ten out of every 1,000,000 novels published between 1960 and\n", " 2010 are written by Pynchon. Of these 10 novels, 9 will test positive as\n", " Pynchon novels. Of the remaining 999,990 novels not written by Pynchon,\n", " approximately 10,000 will test positive as Pynchon novels. Suppose you have a\n", " small sample of novels which have tested positive, how many of the novels are\n", " actually written by Pynchon? The answer is the ratio of true positives to\n", " total positives, or $9 / 10009 \\approx 0.0009$, or 0.09 percent. So the\n", " conclusion is the same. Despite the positive test, the novel in question\n", " remains unlikely to be a Pynchon novel.\n", "```\n", "\n", "This example illustrates the essential features of Bayesian learning: we begin\n", "with a *prior probability*, a description of how likely a novel published\n", "between 1960 and 2010 is a novel by Pynchon, and then update our belief in\n", "light of evidence (a positive test result using a stylometric test), and arrive\n", "at a new *posterior probability*. While we have yet to address the question of\n", "why the use of this *particular* procedure---with its requirement that degree\n", "of belief be expressed as probabilities between 0 and 1---deserves deference,\n", "we now have at least a rudimentary sense about what Bayesian learning involves.\n", "\n", "This extended motivating example prepares us for the case study at the center of this\n", "chapter: the disputed authorship of several historically significant essays by two signers\n", "of the United States Constitution: Alexander Hamilton and James Madison. That this case\n", "study involves, after Roberto Busa's encounters with Thomas Watson, the best known\n", "instance of humanities computing *avant la lettre* will do no harm either. The case study\n", "will, we hope, make abundantly clear the challenge Bayesian inference addresses: the\n", "challenge of describing---to at least ourselves and potentially to other students or\n", "researchers---*how observing a piece of evidence changes our prior beliefs*.\n", "\n", "(sec-intro-probability-probability)=\n", "## Probability\n", "\n", "This section provides a formal introduction to probabilities and their role in\n", "Bayesian inference. Probabilities are the lingua franca of the contemporary\n", "natural and social sciences; their use is not without precedent in the\n", "humanities research either, as the case study in this chapter illustrates.\n", "Authorship attribution is a fitting site for an introduction to using\n", "probabilities in historical research because the task of authorship attribution\n", "has established contours and involves using probabilities to describe beliefs\n", "about events in the distant past---something more typical of historical\n", "research than other research involving the use of probabilities.\n", "\n", "Probabilities are numbers between 0 and 1. For our purposes---and in Bayesian\n", "treatments of probability generally---probabilities express degree of\n", "belief. For example, a climatologist would use a probability (or a distribution\n", "over probabilities) to express their degree of belief in the hypothesis that the\n", "peak temperature in Germany next year will be higher than the previous year. If\n", "one researcher assigns a higher probability to the hypothesis than another\n", "researcher, we say that the first researcher finds (or tends to find) the\n", "hypothesis more *credible*, just as we would likely say in a less formal\n", "discussion.\n", "\n", "For concreteness, we will anchor our discussion of probabilities (degrees of\n", "belief) in a specific case. We will discuss probabilities that have been\n", "expressed by scholars working on a case of disputed authorship. In\n", "1964, Frederick Mosteller (1916–2006) and David Wallace (1928–2017) investigated the authorship of twelve\n", "disputed essays which had been published under a pseudonym in New York\n", "newspapers between 1787 and 1788. The dispute was between two figures well-known in\n", "the history of the colonization of North America by European settlers: Alexander Hamilton and James\n", "Madison. Each claimed to be the sole author of all twelve essays. These twelve essays\n", "were part of a larger body of eightyfive essays known collectively as *The Federalist Papers*.\n", "The essays advocated ratification of the United States Constitution. In\n", "the conclusion of their study, Mosteller and Wallace offer probabilities to\n", "describe their degrees of belief about the authorship of each of the twelve\n", "essays. For example, they report that the probability Federalist No. 62 is\n", "written by Madison is greater than 99.9 percent (odds of 300,000 to 1) while the\n", "probability Federalist No. 55 is written by Madison is 99 percent (100 to 1). In\n", "the present context, it is useful to see how Mosteller expresses their\n", "conclusion: \"Madison is extremely likely, in the sense of degree of belief, to\n", "have written the disputed *Federalist* essays, with the possible exception of\n", "No. 55, and there our evidence is weak; suitable deflated odds are 100 to 1 for\n", "Madison\" ({cite:t}`mosteller1987statistical`, 139).\n", "\n", "It should now be clear what probabilities are. It may not, however, be\n", "entirely clear what problem the *quantitative* description of degree of belief\n", "solves. What need is there, for example, to say that I believe that it will\n", "rain tomorrow with probability 0.8 when I might just as well say, \"I believe it\n", "will likely rain tomorrow\"? Or, to tie this remark to the present case, what is\n", "gained by saying that Madison is very likely the author of Federalist No. 62\n", "versus saying that the odds favoring Madison's authorship of the essay is\n", "300,000 to 1? Those suspicious of quantification *tout court* might consider\n", "the following context as justification for using numbers to make fine\n", "distinctions about degrees of belief. In modern societies, we are familiar\n", "with occasions where one differentiates between various degrees of belief. The\n", "legal system provides a convenient source of examples. Whether you are familiar\n", "with judicial proceedings in a territory making use of a legal system in the\n", "common law tradition (e.g., United States) or a legal system in the civil law\n", "tradition (e.g., France), the idea that evidence supporting conviction must\n", "rise to a certain *level* should be familiar {cite:p}`clermont2002comparative`. If\n", "you are called to serve on a citizen jury (in a common law system) or called to\n", "arbitrate a dispute as a judge (in a civil law system), it is likely not\n", "sufficient to *suspect* that, say, Jessie stole Riley's bike; you will likely\n", "be asked to assess whether you are convinced that Jessie stole Riley's bike\n", "\"beyond a reasonable doubt\". Another classic example of a case where making\n", "fine gradations between degrees of belief would be useful involves the decision\n", "to purchase an insurance contract (e.g., travel insurance). An informed decision requires balancing (1)\n", "the cost of the insurance contract, (2) the cost arising if the event\n", "insured against comes to pass, and (3) the probability that the event will occur.\n", "\n", "\n", "```{note}\n", "Consistent discussion of probabilities in ordinary language is sufficiently\n", "challenging that there are attempts to standardize the terminology. Consider,\n", "for example, the [Intergovernmental Panel on Climate Change's guidance on\n", "addressing\n", "uncertainties](http://www.ipcc.ch/pdf/supporting-material/uncertainty-guidance-note.pdf):\n", "\n", "| Term | Likelihood of the outcome | Odds | Log odds |\n", "|------------------------|---------------------------|----------------|----------|\n", "| Virtually certain | > 99% probability | > 99 (or 99:1) | > 4.6 |\n", "| Very likely | > 90% probability | > 9 | > 2.2 |\n", "| Likely | > 66% probability | > 2 | > 0.7 |\n", "| About as likely as not | 33 to 66% probability | 0.5 to 2 | 0 to 0.7 |\n", "| Unlikely | < 33% probability | < 0.5 | < -0.7 |\n", "| Very unlikely | < 10% probability | < 0.1 | < -2.2 |\n", "| Exceptionally unlikely | < 1% probability | < 0.01 | < -4.6 |\n", "```\n", "\n", "\n", "(sec-intro-probability-degree-of-belief)=\n", "### Probability and degree of belief\n", "\n", "Probabilities are one among many ways of expressing degrees of belief using numbers.\n", "(Forms of the words \"plausibility\" and \"credibility\" will also be used to refer\n", "to \"degree of belief\".) Are they a particularly good way? That is, do\n", "probabilities deserve their status as the canonical means of expressing degrees\n", "of belief? Their characteristic scale (0 to 1) has, at first glance, nothing to\n", "recommend it over alternative continuous scales (e.g., -2 to 2, 1 to 5, or\n", "$-\\infty$ to $\\infty$). Why should we constrain ourselves to this scale when we\n", "might prefer our own particular---and perhaps more personally\n", "familiar---scale? Presently, reviewing products (books,\n", "movies, restaurants, hotels, etc.) in terms of a scale of 1 to 5 \"stars\" is\n", "extremely common. A scale of 0 to 1 is rarely used.\n", "\n", "A little reflection will show that the scale used by probabilities has a few\n", "mathematical properties that align well with a set of minimal requirements for\n", "deliberation involving degrees of belief. For example, the existence of a lower\n", "bound, $0$, is consistent with the idea that, if we believe an event is\n", "impossible (e.g., an event which has already failed to occur), there can be\n", "no event in which we have a lower degree of belief. Similarly, the\n", "existence of an upper bound, $1$, is consistent with the assumption that if we\n", "believe an event is certain (e.g., an event which has already occurred), there\n", "can be no event which is more plausible. Any bounded interval has this\n", "property; other intervals, such as zero to infinity, open\n", "intervals, or the natural numbers, do not.\n", "\n", "```{margin}\n", "Formally the\n", "axioms of beliefs are more general than the axioms of probability. If, however,\n", "you use the axioms of beliefs and a 0 to 1 scale, you satisfy both the axioms\n", "of beliefs and the axioms of probability ({cite:t}`hoff2009first`, 13--14).\n", "```\n", "The use of probabilities is also associated with a minimal set of rules, the\n", "\"axioms of belief\" or the \"axioms of probability\", which must be followed when\n", "considering the aggregate probability of the conjunction or disjunction of\n", "events ({cite:t}`hoff2009first`, 13--14; {cite:t}`casella2001statistical`, 7--10). (The conjunction of events A and B is often expressed as \"A or B\", the disjunction as \"A and B\".)\n", "These rules are often used to derive other rules. For example, from the rule that all probabilities must be between 0 and 1, we can conclude that the probability of an aggregate event $C$ that either event $A$ occurs or\n", "event $B$ occurs (often written as $C = A \\cup B$, where $\\cup$ denotes the\n", "union of elements in two sets) must also lie between 0 and 1.\n", "\n", "\n", "\n", "```{note}\n", "The axioms of probability, despite the appearance of the superficially\n", "demanding term \"axiom\", only minimally constrain deliberations about events.\n", "Adhering to them is neither a mark of individual nor general obedience to a\n", "broader (universalizing) set of epistemological dispositions (e.g.,\n", "\"rationality\"). The axioms demand very little. They demand, in essence, that\n", "we be willing to both bound and order our degrees of belief. They do not say\n", "anything about what our initial beliefs should be. If we are uncomfortable\n", "with, for example, ever saying ourselves (or entertaining someone else's\n", "saying) that one event is more likely than another event (e.g., it is more\n", "likely that the writer of *The Merchant of Venice* was human than non-human),\n", "then the use of probabilities will have to proceed, if it proceeds at all, as\n", "a thought experiment. (A pessimistic view of the history of scientific claims\n", "of knowledge might figure in such a refusal to compare or talk in terms of\n", "degrees of belief.) Such a thought experiment would, however, find ample\n", "pragmatic justification: observational evidence and testimony in favor of the\n", "utility of probabilities is found in a range of practical endeavours, such as\n", "engineering, urban planning, and environmental science. And we\n", "need not explicitly endorse the axioms of probability in order to use them or\n", "to recommend their use.\n", "```\n", "\n", "\n", "```{margin}\n", "[Grinstead and Snell's Introduction to\n", "Probability](https://math.dartmouth.edu/~prob/prob/prob.pdf) is excellent and freely \n", " available online. \n", "```\n", "When we use probabilities to represent degrees of belief, we use the following standard\n", "notation. The degree of belief that an event $A$ occurs is $\\Pr(A)$. Events may be\n", "aggregated by conjunction or disjunction: $A \\textrm{ or } B$ is itself an event, as is $A\n", "\\textrm{ and } B$. The degree of belief that $A$ occurs given that event $B$ has already\n", "occurred is written $\\Pr(A|B)$. For a formal definition of \"event\" in this setting and a\n", "thorough treatment of probability, readers are encouraged to consult a text dedicated to\n", "probability, such as {cite:t}`grinstead2012introduction` or\n", "{cite:t}`hacking2001introduction`.\n", "\n", "```{margin}\n", "The axioms of\n", "beliefs are satisfied, with very minor adjustments, by $\\Pr$, considered as a\n", "belief function, which respects the axioms of probability. Interested readers\n", "are encouraged to review {cite:t}`jaynes2003probability` or {cite:t}`kadane2011principles`.\n", "```\n", "Using this notation we can state the axioms of probability:\n", "\n", "1. $0 \\le \\Pr(A) \\le 1$.\n", "2. $\\Pr(A \\textrm{ or } B) = \\Pr(A) + \\Pr(B)$ if $A$ and $B$ are mutually exclusive events.\n", "3. $\\Pr(A \\textrm{ and } B) = \\Pr(B) \\Pr(A|B)$.\n", "\n", "```{margin}\n", "E. T. Jaynes quotes\n", "Laplace approvingly, \"Probability theory is nothing but common sense reduced to\n", "calculation\" ({cite:t}`jaynes2003probability`, 24).\n", "```\n", "Each of these axioms can be translated into reasonable constraints on the ways\n", "in which we are allowed to describe our degrees of belief. An argument for axiom 1 has been\n", "offered above. Axiom 2 implies that our degree of belief that one event in a\n", "collection of possible events will occur should not decrease if the set of\n", "possible events expands. (My degree of belief in rain occurring tomorrow must be less than or equal to my belief rain *or* snow will occur tomorrow.)\n", "Axiom 3 requires that the probability of a complex\n", "event be calculable when the event is decomposed into component \"parts\". The\n", "rough idea is this: we should be free to reason about the probability of $A$\n", "and $B$ both occurring by first considering the probability that $B$ occurs and\n", "then considering the probability that $A$ occurs given that $B$ has occurred.\n", "\n", "From the axioms of probability follows a rule, Bayes's rule, which delivers a method of\n", "learning from observation. Bayes's rule offers us a useful\n", "prescription for how our degree of belief in a hypothesis should change\n", "(or \"update\") after we observe evidence bearing on the hypothesis. For example, we\n", "can use Bayes's rule to calculate how our belief that Madison is the author of\n", "Federalist No. 51 (a hypothesis) should change after observing the number of\n", "times the word *upon* occurs in No. 51 (evidence). (*Upon* occurs much more\n", "reliably in Hamilton's previous writings than in Madison's.)\n", "\n", "Bayes's rule is remarkably general. If we can describe our *prior* belief about a\n", "hypothesis $H$ as $\\Pr(H)$ (e.g., the hypothesis that Madison is the author of\n", "Federalist No. 51, $\\Pr(\\textrm{Madison})$), and describe how likely it would\n", "be to observe a piece of evidence $E$ given that the hypothesis holds,\n", "$\\Pr(E|H)$ (e.g., the likelihood of observing *upon* given the hypothesis of\n", "Madison's authorship, $\\Pr(\\textrm{upon}|\\textrm{Madison})$), Bayes's rule tells\n", "us how to \"update\" our beliefs about the hypothesis to arrive at our *posterior*\n", "degree of belief in the hypothesis, $\\Pr(H|E)$\n", "(e.g., $\\Pr(\\textrm{Madison}|\\textrm{upon})$), in the event that we indeed observe\n", "the specified piece of evidence. Formally, in a situation where there are $K$\n", "competing hypotheses, Bayes' rule reads as follows,\n", "\n", "\\begin{equation}\\label{eq:bayes-rule}\n", "\\Pr(H_k|E) = \\frac{\\Pr(E|H_k)\\Pr(H_k)}{\\sum_{j=1}^K \\Pr(E|H_j)\\Pr(H_j)},\n", "\\end{equation}\n", "\n", "where $\\Pr(H_k|E)$ is the probability of hypothesis $k$ ($H_k \\in \\mathcal{H}$) given that\n", "evidence $E$ is observed. Here we assume that the hypotheses are mutually exclusive and\n", "exhaust the space of possible hypotheses. In symbols we would write that the hypotheses\n", "$H_k, k = 1, \\ldots, K$ partition the hypothesis space: $\\cup_{k=1}^K H_k = \\mathcal{H}$\n", "and $\\cap_{k=1}^K H_k = \\emptyset$. Bayes' rule follows from the axioms of probability. A\n", "derivation of Bayes' rule is provided in an appendix to this\n", "chapter (Appendix {ref}`appendix-intro-probability-bayes-rule`). A fuller appreciation of\n", "Bayes's rule may be gained by considering its application to questions (that is,\n", "hypotheses) of interest. We turn to this task in the next section.\n", "\n", "(sec-intro-probability-bayes-rule-authorship-attribution)=\n", "## Example: Bayes's Rule and Authorship Attribution\n", "\n", "In the case of authorship attribution of the disputed essays in *The Federalist Papers*, the\n", "space of hypotheses $\\mathcal{H}$ is easy to reason about: a disputed essay is\n", "written either by Hamilton or by Madison. There is also not much difficulty in\n", "settling on an initial \"prior\" hypothesis: the author of a disputed\n", "*Federalist* essay---we should be indifferent to the particular (disputed)\n", "essay---is as likely to be Hamilton as it is to be Madison. That is, the\n", "probability that a disputed essay is written by Hamilton (prior to observing\n", "any information about the essay) is equal to 0.5, the same probability which we\n", "would associate with the probability that the disputed essay is written by\n", "Madison. In symbols, if we use $H$ to indicate the event of Hamilton being the\n", "author then $\\{ H, \\neg H\\}$ partitions the space of possible hypotheses.\n", "$\\Pr(H) = \\Pr(\\neg H) = 0.5$ expresses our initial indifference between the\n", "hypothesis that Hamilton is the author and the hypothesis that Madison (i.e.,\n", "not Hamilton) is the author.\n", "\n", "For this example, the evidence, given a hypothesis about the author, is the\n", "presence or absence of one or more occurrences of the word *upon* in the essay.\n", "This word is among several words which Mosteller and Wallace identify as both\n", "having a distinctive pattern in the essays of known authorship and being the\n", "kind of frequent \"function\" word widely accepted as having no strong connection\n", "between the subject matter addressed by a piece of writing. These function words\n", "are considered as promising subject-independent markers of authorial\n", "\"style\" (chapter {ref}`chp:stylometry` delves deeper into the added value of\n", "function words in authorship studies). The code blocks below will construct a\n", "table describing the frequency with which *upon* occurs across 51 essays by\n", "Hamilton and 36 essays by Madison.\n", "\n", "```{note}\n", "The dataset used here includes all the\n", "disputed essays, all the essays in *The Federalist Papers*, and a number of additional\n", "(non-disputed) essays. Seventy-seven essays in *The Federalist Papers* were published in *The Independent\n", "Journal* and *the New York Packet* between October 1787 and August 1788. Twelve\n", "essays are disputed. John Jay wrote 5 essays, Hamilton wrote 43 essays, and\n", "Madison wrote 14 essays. Three essays were jointly written by Hamilton and\n", "Madison; their authorship is not disputed ({cite:t}`mosteller1987statistical`,\n", "132). Additional essays (also grouped under the heading of *The Federalist Papers*)\n", "were published by Hamilton after the 77 serialized essays. In order to\n", "have a better sense of the variability in Madison's style, additional essays by\n", "Madison from the period were included in the analysis. The dataset used here\n", "includes all these essays. A flawless reproduction of the dataset from the\n", "original sources that Mosteller and Wallace used does not yet exist. Creating\n", "one would be an invaluable service.\n", "```\n", "\n", "The first code block introduces *The Federalist Papers* essays' word frequencies,\n", "stored in a CSV file ``federalist-papers.csv``, by randomly sampling several\n", "essays (using ``DataFrame``'s ``sample()`` method) and displaying the frequency\n", "of several familiar words. The dataset containing the word frequencies is\n", "organized in a traditional manner with each row corresponding to an essay and\n", "the columns named after (lower cased) elements in the vocabulary." ] }, { "cell_type": "code", "execution_count": 3, "id": "566ff89f", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
uponthestateenoughwhile
6821421000
3662512500
743104200
630290600
400294600
5422041600
\n", "
" ], "text/org": [ "| | upon | the | state | enough | while |\n", "|----+--------+-------+---------+----------+---------|\n", "| 68 | 2 | 142 | 10 | 0 | 0 |\n", "| 36 | 6 | 251 | 25 | 0 | 0 |\n", "| 74 | 3 | 104 | 2 | 0 | 0 |\n", "| 63 | 0 | 290 | 6 | 0 | 0 |\n", "| 40 | 0 | 294 | 6 | 0 | 0 |\n", "| 54 | 2 | 204 | 16 | 0 | 0 |" ], "text/plain": [ " upon the state enough while\n", "68 2 142 10 0 0\n", "36 6 251 25 0 0\n", "74 3 104 2 0 0\n", "63 0 290 6 0 0\n", "40 0 294 6 0 0\n", "54 2 204 16 0 0" ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import pandas as pd\n", "import numpy as np; np.random.seed(1) # fix a seed for reproducible random sampling\n", "\n", "# only show counts for these words:\n", "words_of_interest = ['upon', 'the', 'state', 'enough', 'while']\n", "df = pd.read_csv('data/federalist-papers.csv', index_col=0)\n", "df[words_of_interest].sample(6)" ] }, { "cell_type": "markdown", "id": "2f7e679b", "metadata": {}, "source": [ "In order to show the frequency of *upon* in essays with known authors we first\n", "verify that the dataset aligns with our expectations. Are there, as we\n", "anticipate, twelve disputed essays? The dataset includes a column with the name\n", "\"AUTHOR\" which records received opinion about the authorship of the essays\n", "(that is, before Mosteller and Wallace's research). We can identify the rows\n", "associated with disputed essays by finding those rows whose ``AUTHOR`` is\n", "\"HAMILTON OR MADISON\" (i.e., disputed):" ] }, { "cell_type": "code", "execution_count": 4, "id": "22bffd7b", "metadata": {}, "outputs": [], "source": [ "# values associated with the column 'AUTHOR' are one of the following:\n", "# {'HAMILTON', 'MADISON', 'JAY', 'HAMILTON OR MADISON',\n", "# 'HAMILTON AND MADISON'}\n", "# essays with the author 'HAMILTON OR MADISON' are the 12 disputed essays.\n", "disputed_essays = df[df['AUTHOR'] == 'HAMILTON OR MADISON'].index\n", "assert len(disputed_essays) == 12 # there are twelve disputed essays\n", "# numbers widely used to identify the essays\n", "assert set(disputed_essays) == {49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 62, 63}" ] }, { "cell_type": "markdown", "id": "d5f6c635", "metadata": {}, "source": [ "Now we gather texts where authorship is known by locating rows in our dataset\n", "where the value associated with the \"AUTHOR\" column is either \"HAMILTON\" or\n", "\"MADISON\" (indicating exclusive authorship). We make use of the ``Series.isin()`` method\n", "to identify which elements in a ``Series`` are members of a provided\n", "sequence. The ``isin()`` method returns a new ``Series`` of ``True`` and ``False``\n", "values." ] }, { "cell_type": "code", "execution_count": 5, "id": "cff4646e", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "HAMILTON 51\n", "MADISON 36\n", "Name: AUTHOR, dtype: int64\n" ] } ], "source": [ "# gather essays with known authorship: the undisputed essays of\n", "# Madison and Hamilton\n", "df_known = df.loc[df['AUTHOR'].isin(('HAMILTON', 'MADISON'))]\n", "print(df_known['AUTHOR'].value_counts())" ] }, { "cell_type": "markdown", "id": "d8a06171", "metadata": {}, "source": [ "The frequency of *upon* in each essay is recorded in the column with the label\n", "\"upon\", as in the familiar vector space representation (cf. \n", "{ref}`chp-vector-space-model`). Overlapping histograms communicate\n", "visually the difference in the usage of *upon* in the essays of known\n", "authorship. The following block of code creates such a plot and uses the\n", "``DataFrame`` method ``groupby()`` to generate, in effect, two ``DataFrame``s with\n", "uniform \"AUTHOR\" values (i.e., either \"HAMILTON\" or \"MADISON\" but not both):\n", "\n", "\n", "" ] }, { "cell_type": "code", "execution_count": 6, "id": "e272bf8a", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans.\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "findfont: Generic family 'sans-serif' not found because none of the following families were found: \"Roboto Condensed Regular\"\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "filenames": { "image/png": "/Users/folgert/projects/hda/_build/jupyter_execute/intro-probability/notebook_11_2.png" } }, "output_type": "display_data" } ], "source": [ "df_known.groupby('AUTHOR')['upon'].plot.hist(\n", " rwidth=0.9, alpha=0.6, range=(0, 22), legend=True);" ] }, { "cell_type": "markdown", "id": "3c67f2f8", "metadata": {}, "source": [ "\n", "\n", "The difference, indeed, is dramatic enough that the visual aid of the histogram is not strictly necessary. Hamilton uses *upon* far more\n", "than Madison. We can gain a sense of this by comparing any number of summary statistics\n", "(see {ref}`chp-statistics-essentials`), including the maximum frequency of *upon*, the\n", "mean frequency, or the median (50th percentile). These summary statistics are conveniently\n", "produced by the ``DataFrame`` method ``describe()``:" ] }, { "cell_type": "code", "execution_count": 7, "id": "542393d9", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
countmeanstdmin25%50%75%max
AUTHOR
HAMILTON51.07.3333334.0083252.04.06.010.0020.0
MADISON36.01.2500001.5743480.00.00.52.255.0
\n", "
" ], "text/org": [ "| AUTHOR | count | mean | std | min | 25% | 50% | 75% | max |\n", "|----------+---------+---------+---------+-------+-------+-------+-------+-------|\n", "| HAMILTON | 51 | 7.33333 | 4.00832 | 2 | 4 | 6 | 10 | 20 |\n", "| MADISON | 36 | 1.25 | 1.57435 | 0 | 0 | 0.5 | 2.25 | 5 |" ], "text/plain": [ " count mean std min 25% 50% 75% max\n", "AUTHOR \n", "HAMILTON 51.0 7.333333 4.008325 2.0 4.0 6.0 10.00 20.0\n", "MADISON 36.0 1.250000 1.574348 0.0 0.0 0.5 2.25 5.0" ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df_known.groupby('AUTHOR')['upon'].describe()" ] }, { "cell_type": "code", "execution_count": 8, "id": "21f5cbd9", "metadata": { "tags": [ "remove-cell" ] }, "outputs": [], "source": [ "np.testing.assert_array_almost_equal(\n", " df_known.groupby('AUTHOR')['upon'].mean().values, [7.33333333, 1.25])" ] }, { "cell_type": "markdown", "id": "c537d01b", "metadata": {}, "source": [ "In this particular case, we get an even clearer picture of the difference\n", "between the two writers' use of *upon* by considering the proportion of essays\n", "in which *upon* appears at all (i.e., one or more times). In our sample,\n", "Hamilton always uses *upon* whereas we are as likely as not to observe it in\n", "an essay by Madison." ] }, { "cell_type": "code", "execution_count": 9, "id": "8da270c9", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "AUTHOR\n", "HAMILTON 1.0\n", "MADISON 0.5\n", "Name: upon, dtype: float64\n" ] } ], "source": [ "# The expression below applies `mean` to a sequence of binary observations\n", "# to get a proportion. For example,\n", "# np.mean([False, False, True]) == np.mean([0, 0, 1]) == 1/3\n", "proportions = df_known.groupby('AUTHOR')['upon'].apply(\n", " lambda upon_counts: (upon_counts > 0).mean())\n", "print(proportions)" ] }, { "cell_type": "markdown", "id": "599f2150", "metadata": {}, "source": [ "In the preceding block we make use of the methods ``groupby()`` (of ``DataFrame``)\n", "and ``apply()`` (of ``SeriesGroupBy``)---cf. section\n", "{ref}`sec-working-with-data-turnover-in-naming-practices`. Recall that the ``SeriesGroupBy`` method ``apply()``\n", "works in the following manner: provided a single, callable argument (e.g., a\n", "function or a lambda expression), ``apply()`` calls the argument with the discrete\n", "individual series generated by the ``groupby()`` method. In this case, ``apply()``\n", "first applies the ``lambda`` to the sequence of *upon* counts associated with\n", "Hamilton and then to the sequence of counts associated with Madison. The\n", "``lambda`` expression itself makes use of a common pattern in scientific\n", "computing: it calculates the average of a sequence of binary (0 or 1)\n", "observations. When the value 1 denotes the presence of a trait and 0 denotes its\n", "absence---as is the case here---the mean will yield the proportion of\n", "observations in which the trait is present. For example, if we were studying the\n", "presence of a specific feature in a set of five essays, we might evaluate\n", "``np.mean([0, 0, 1, 1, 0])`` and arrive at an answer of $\\frac{2}{5} = 0.4$.\n", "\n", "With the heavy lifting behind us, we can create a bar chart which shows the\n", "proportion of essays (of known authorship) which contain *upon*, for each\n", "author." ] }, { "cell_type": "code", "execution_count": 10, "id": "f3ff36a5", "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "filenames": { "image/png": "/Users/folgert/projects/hda/_build/jupyter_execute/intro-probability/notebook_18_0.png" } }, "output_type": "display_data" } ], "source": [ "proportions.plot.bar(rot=0);" ] }, { "cell_type": "markdown", "id": "cc1a5338", "metadata": {}, "source": [ "" ] }, { "cell_type": "code", "execution_count": 11, "id": "bfb0e6ff", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "AUTHOR\n", "HAMILTON 1.0\n", "MADISON 0.5\n", "Name: upon, dtype: float64" ] }, "execution_count": 11, "metadata": {}, "output_type": "execute_result" } ], "source": [ "proportions" ] }, { "cell_type": "markdown", "id": "f748ab6a", "metadata": {}, "source": [ "Hamilton uses *upon* more often and more consistently than Madison. While the\n", "word *upon* appears in every one of the 51 essays by Hamilton (100%), the word\n", "occurs in only 18 of the 36 essays by Madison (50%). Observing this\n", "difference, we should conclude that observing the word *upon* one or more times\n", "in a disputed essay is more consistent with the hypothesis that the author is\n", "Hamilton than with the hypothesis that Madison is the author. Precisely how\n", "much more consistent is something which must be determined.\n", "\n", "Observing the word *upon* in a disputed essay is hardly decisive evidence that\n", "Hamilton wrote the essay; Madison uses the word *upon* occasionally and, even if\n", "he never was observed to have used the word, we need to keep in mind the wide\n", "range of unlikely events that could have added an *upon*\n", "to a Madison-authored essay, such as the insertion of an *upon* by an\n", "editor. These considerations will inform a precise description of the\n", "probability of the evidence, something we need if we want to use Bayes's rule and\n", "update our belief about the authorship of a disputed essay.\n", "\n", "The following is a description of the probability of the evidence (observing\n", "the word *upon* one or more times in a disputed essay) given the hypothesis\n", "that Hamilton is the author. Suppose we think that observing *upon* is far from\n", "decisive. Recalling that probabilities are just descriptions of degree of\n", "belief (scaled to lie between 0 and 1 and following the axioms of probability),\n", "we might say that if Hamilton wrote the disputed essay, the probability of it\n", "containing one or more occurrences of the word *upon* is 0.7. In symbols, we\n", "would write $\\Pr(E|H) = 0.7$. We're using $\\Pr(E|H)$ to denote the probability\n", "of observing the word *upon* in an unobserved essay given that Hamilton is the\n", "author.\n", "\n", "Now let's consider Madison. Looking at how often Madison uses the word *upon*,\n", "we anticipate that it is as likely as not that he will use the word (one or more\n", "times) if the disputed essay is indeed written by him. Setting $\\Pr(E|\\neg H) = 0.5$\n", "seems reasonable. Suppose that we are now presented with one of the disputed\n", "essays and we observe that the word *upon* does indeed occur one or more\n", "times. Bayes's rule tells us how our belief should change after observing this\n", "piece of evidence:\n", "\n", "\\begin{equation}\\label{eq:bayes-after-evidence}\n", "\\Pr(H_j|E) = \\frac{\\Pr(E|H_j)\\Pr(H_j)}{\\sum_{k=1}^K \\Pr(E|H_k)\\Pr(H_k)}\n", "\\end{equation}\n", "\n", "Since the hypothesis space is partitioned by $H$ and $\\neg H$ (\"not H\"), the formula can\n", "be rewritten. As we only have two possible authors (Hamilton or Madison) we\n", "can write $H$ for the hypothesis that Hamilton wrote the disputed essay and\n", "$\\neg H$ for the hypothesis that Madison (i.e., not Hamilton) wrote\n", "the disputed essay. And since we have settled on values for all the terms on\n", "the right-hand side of the equation we can complete the calculation:\n", "\n", "$$\\begin{aligned}\n", "\\Pr(H|E) &= \\frac{\\Pr(E|H)\\Pr(H)}{\\Pr(E|H)\\Pr(H) + \\Pr(E|\\neg H)\\Pr(\\neg H)} \\\\\n", " &= \\frac{(0.7)(0.5)}{(0.7)(0.5) + (0.5)(0.5)} \\approx 0.58 \\\\\n", "\\end{aligned}$$\n", "\n", "It makes sense that our degree of belief in Hamilton being the author increases\n", "from 0.5 to 0.58 in this case: Hamilton uses *upon* far more than Madison, so\n", "seeing this word in a disputed essay should increase (or, at least, certainly\n", "not decrease) the plausibility of the claim that Hamilton is the author. The\n", "precise degree to which it changes is what Bayes's rule contributes.\n", "\n", "(sec-intro-probability-random-variables)=\n", "### Random variables and probability distributions\n", "\n", "In the previous example, we simplified the analysis by only concerning ourselves with\n", "whether or not a word occurred one or more times. In this specific case the simplification\n", "was not consequential, since we were dealing with a case where the presence or absence of\n", "a word was distinctive. Assessing the distinctiveness of word usage in cases where the\n", "word is common (i.e., it typically occurs at least once) requires the use of a discrete\n", "probability distribution. One distribution which will allow us to model the difference\n", "with which authors use words such as *by* and *whilst* is the negative\n", "binomial distribution. When working with models of individual word frequencies in\n", "text documents, the negative binomial distribution is almost always a better choice than\n", "the binomial distribution.\n", "\n", "```{margin}\n", "{cite:t}`simon1955class` called attention to this property of the\n", "negative binomial distribution. The relevance of this property to Mosteller and\n", "Wallace's choice is discussed in section 3.1 of {cite:t}`airoldi2006who`.\n", "```\n", "The key feature of texts that the\n", "negative binomial distribution captures and that the binomial\n", "distribution does not is the *burstiness* or *contagiousness* of uncommon words:\n", "if an uncommon word, such as a location name or person's name, appears in a\n", "document, the probability of seeing the word a second time should go up\n", "{cite:p}`church1995poisson`.\n", "\n", "Examining the frequency with which Madison and Hamilton use the word *by*, it\n", "is easy to detect a pattern. Based on the sample we have, Madison reliably uses\n", "the word *by* more often than Hamilton does: Madison typically uses the word 13\n", "times per 1,000 words; Hamilton typically uses it about 6 times per 1,000\n", "words. To reproduce these rate calculations, we first scale all word frequency\n", "observations into rate per 1,000 words. After doing this, we exclude essays not\n", "written by Hamilton or Madison and then display the average rate of *by*." ] }, { "cell_type": "code", "execution_count": 12, "id": "6f6f1f2f", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
countmeanstdmin25%50%75%max
AUTHOR
HAMILTON51.07.0196084.9254041.05.06.09.034.0
MADISON36.012.9166674.0944215.010.012.515.023.0
\n", "
" ], "text/org": [ "| AUTHOR | count | mean | std | min | 25% | 50% | 75% | max |\n", "|----------+---------+----------+---------+-------+-------+-------+-------+-------|\n", "| HAMILTON | 51 | 7.01961 | 4.9254 | 1 | 5 | 6 | 9 | 34 |\n", "| MADISON | 36 | 12.9167 | 4.09442 | 5 | 10 | 12.5 | 15 | 23 |" ], "text/plain": [ " count mean std min 25% 50% 75% max\n", "AUTHOR \n", "HAMILTON 51.0 7.019608 4.925404 1.0 5.0 6.0 9.0 34.0\n", "MADISON 36.0 12.916667 4.094421 5.0 10.0 12.5 15.0 23.0" ] }, "execution_count": 12, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df = pd.read_csv('data/federalist-papers.csv', index_col=0)\n", "author = df['AUTHOR'] # save a copy of the author column\n", "df = df.drop('AUTHOR', axis=1) # remove the author column\n", "df = df.divide(df.sum(axis=0)) # rate per 1 word\n", "df *= 1000 # transform from rate per 1 word to rate per 1,000 words\n", "df = df.round() # round to nearest integer\n", "df['AUTHOR'] = author # put author column back\n", "df_known = df[df['AUTHOR'].isin({'HAMILTON', 'MADISON'})]\n", "\n", "df_known.groupby('AUTHOR')['by'].describe()" ] }, { "cell_type": "markdown", "id": "9854899f", "metadata": {}, "source": [ "We can visualize the difference in the authors' use of *by* with two overlapping\n", "histograms:\n", "\n", "" ] }, { "cell_type": "code", "execution_count": 13, "id": "641c2d2c", "metadata": {}, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAABWEAAAPdCAYAAADrsapyAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/NK7nSAAAACXBIWXMAAB7CAAAewgFu0HU+AABv10lEQVR4nOzdedSVdb3//9cGVERUBEUZxFlzVtCch8Ahc0gtpzyCQ8661LQyFdHSsqOZnjQVUzCH1BxyyHnCCZyQtDRNEQU0cUICBAT27w9+7C8kCDfsD/eNPh5rsc61976uz/XeHtY66zzXh2tXqtVqNQAAAAAAFNGssQcAAAAAAPgqE2EBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmG/xjbZZJN07tw5m2yySWOPAgAAAABfWS0aewAaz7///e+MHDmysccAAAAAgK80O2EBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAApq0dgDAAAAADQFfV8e09gjNIoj1l9qgd2rUqnUjqvVavHrzj///PzkJz+pvb755puzzz77zNW1w4YNyyqrrDLTez/96U9z3nnnzdX1V155ZY444oiZ3nv00Uez/fbbf+Hc7bffPgMGDJjvc2Y187zq1atX+vfvP8vPqtVq7rnnnvz1r3/Nk08+mffffz+ffPJJ2rRpkxVWWCFbbbVVdt1113znO99Js2Zz3gN61lln5eyzz6693nzzzTNw4MA5XnfwwQfnmmuuSZI8+OCD2WGHHebuyzUCO2EBAAAA+Eq6+uqrv/R1Q/3xj3/MlClT5uneXxWPPfZYunbtmt122y2XXXZZXn755YwaNSqff/55Pvjgg7z88su5/PLLs/vuu2fjjTfOo48+2uB7DBo0KLfffnuB6RuPnbAAAAAAfOU89dRT+ec//znTew888EBGjBiRzp07N2itFi1aZPLkyXnvvfdy7733ZrfddvvS81999dUMGjRopmsXhPbt239pvPz73/+e3r17J0nWXXfdnHPOObM9t0uXLl9474orrsixxx5bC9HLLrts9txzz3Tt2jXt2rXLxx9/nBdffDF/+ctfMmrUqLz00kvZcccdc+mll+bII49s0Hc5/fTTs8cee6R58+YNuq6pEmEBAAAA+Mq56qqraseHHHJI+vXrl6lTp6Z///4544wzGrRWt27dMmzYsLz//vvp16/fHCPs9F2wiy66aLp375777ruv4V9gHrRq1Sp77rnnbD9v06ZN7Xh6QJ1bN954Y4466qja6xNOOCHnnHNOWrdu/YVzL7zwwvTu3Tu//e1vM2XKlBx11FFp06ZN9ttvvzneZ4kllsi4cePy6quvpn///jnssMPmesamzOMIAAAAAPhK+c9//pObb745SbLGGmvkoosuyuKLL54k6devX4OeK5tM28160EEHJUnuuuuufPjhh7M9d/Lkybn22muTJN/97nfTrl27efkKTcqwYcNmer7tueeem4suumiWATaZFlIvvPDCnHvuubX3Dj/88AwbNmyO9zr22GPTosW0faNnnXVWJkyYMH/DNxEiLAAAAABfKTfddFPGjRuXJDnooIOy1FJLZa+99kqSDB06NI899liD1zz00EOTJJ9//nktss7KX//617z//vszXbOwO++88/Kf//wnSbLDDjvktNNOm6vrTjvttPTo0SPJtDD+61//eo7XrLHGGvnhD3+YJBkxYkR+97vfzePUTYsICwAAAMBXyvRHEVQqldoO1l69en3h84ZYe+21s/nmmyeZtpt2dqY/iqBTp07ZaaedGnyfpmb06NG55ppraq9/8YtfNOj6n//857Xj/v3759NPP53jNX369EmrVq2SJL/61a8yevToBt2zKRJhAQAAAPjKeOWVV2o/irXddttl5ZVXTjJtB2enTp2SJLfddttcxcD/Nn1n68svv5znn3/+C5+///77ueeee5JMi77Nmi386e3xxx+vPRJgzTXXrIXoubXllltmjTXWSJJMmDAhTzzxxByvWWGFFXLSSSclST755JOcd955DZy66Vn4/yYAAAAAwP9vxl2uM+5+bdasWW1X7GeffZYbbrihwWvvv//+tR2a03e8zuiPf/xjJk+enGTaj4F9FTz11FO146222mqe1thyyy1rx08++eRcXfPjH/+49jzd//u//8u77747T/duKkRYAAAAABa4SqUy13/m1ozPa23VqlW+//3vz/T5jFF2VhF1TpZccsnss88+SZI//elPX/jRqOmPKdh2222z+uqrN3j9pmjEiBG147XWWmue1vjGN75ROx45cuRcXbP00kvXnj372Wef5ayzzpqnezcVIiwAAAAAXwl33nlnPvjggyTJ3nvvndatW8/0+Te+8Y1885vfTJI8//zzeemllxp8j+mPJBg9enRuv/322vsDBw7Mq6++OtM5XwUff/xx7bhNmzbztMaM13300Udzfd2xxx6bLl26JJkWzV977bV5un9T0KKxBwAAAADg62fGgDkne+2111ydN+Pu1hl3vc6oV69eefbZZ5NMe3TBxRdfPNdzJP9vl+sbb7yRq6++OgcccMBM915yySW/sAOXebPYYovl7LPPziGHHJIpU6bk9NNPzy233NLYY80TERYAAACABW7PPfes63ojR47M/fffnyTp3LlzunfvPsvz9t9//5x00kmZNGlSrr/++px//vlZdNFFG3SvQw45JKeffnoeeeSRvPPOO1l22WVz0003JUn23XffLLHEEvP3ZZqQtm3b1o5Hjx49T2vMeN3057zOrZ49e+aCCy7IP/7xj9x666159tlna7uZFyYeRwAAAADAQq9///6ZMmVKkuR//ud/0qzZrLNX27Zts/vuuyeZ9k/j//KXvzT4Xr169Urz5s0zderU9O/fP3/+85/zn//8J8lX61EEybSgPd28Pg7gn//8Z+24U6dODbq2WbNm+eUvf1l7feqpp87TDI1NhAUAAABgoVatVmd6FMF55533pT/0deutt9bOnZcf6OrUqVN22mmnJNPi71VXXZVk2jNnt9xyy/n8Nk3LVlttVTt+6qmn5mmNgQMH1o633nrrBl+/xx571OZ49NFHazueFyYeRwA0eX1fHtPYIzALR6y/VGOPAAAAkCR57LHHMnTo0Hm69sEHH8zw4cOz4oorNui6Qw89NPfee2/eeuutvPXWW0mmPabgq2bbbbdNy5YtM2HChLz++ut55plnstlmm8319QMHDszrr7+eJGnZsmW22WabeZrjvPPOq1176qmn1iL4wkKEBQAAAGChNn0napJ873vfy3rrrTfHa55++uk8+OCDtUcK9O7du0H33GOPPbLsssvmww8/TJI0b948Bx10UMMGXwi0adMmPXv2TN++fZMkZ555ZoN2ovbp06d2fMghh2TppZeepzm23nrr7L777rnrrrsyZMiQ/OlPf5qndRqLCAsAAADAQmv06NG1xwu0aNEil112WZZbbrk5XvfSSy9lww03TJL069cvZ5xxRiqVylzfd9FFF83JJ59ce6bsJptskg4dOjT8CywETj311Nxwww0ZO3ZsHnjggfz617/OT3/60zle9+tf/zoPPvhgkmTJJZecq2u+zC9/+cv89a9/zdSpU9O7d+9svvnm87XeguSZsAAAAAAstG644YZMmDAhSbLLLrvMVYBNkg022CAbbbRRkuStt97Ko48+2uB7n3rqqRk0aFAGDRqUSy65pMHXLyxWWWWV2k7YZNr3PvnkkzNu3LhZnj9+/Pj8+Mc/nulHtK688sqstNJK8zXHeuutV9ttPHTo0Nxxxx3ztd6CZCcsAAAAAAutGR9F0LNnzwZd27NnzwwZMqS2Tvfu3es5WjFXXXVVHnroobk694wzzkjLli3n+54HHHBAPv300xx33HGZMmVKLrzwwlx77bXZa6+90rVr17Rt2zYff/xxhgwZkttuuy2jRo1KMu0xDZdcckn222+/+Z4hSX7+85/nxhtvzMSJE2cbgZsiERYAAACAhdKQIUMyePDgJMkyyyyT3XffvUHXH3jggfnJT36SyZMn57bbbsvo0aPTpk2bApPW13XXXTfX555yyil1ibBJctRRR2WttdbKSSedlL/97W/54IMPZtoh+9822GCDXHTRRfnWt75Vl/snSZcuXXLMMcfkt7/9bd3WXBA8jgAAAACAhdKMu2D322+/LLbYYg26vn379vn2t7+dJJkwYUJuuOGGus73VfStb30rL774Yu6+++4cddRRWW+99bLsssumRYsWWXbZZbPuuuvmiCOOyJ133pkXX3yxrgF2utNPPz1LLbVU3dctqVKtVquNPQSNo3Pnzhk5cmQ6deqUESNGNPY4MFt9Xx7T2CMwC0esv3D9HzwAAABoLHbCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAU1KKxBwAAAABoCj59+sjGHqFRLL3lFQvsXpVKZabX++23X2688ca5uvbBBx/MTjvtNNN7/fr1y8EHHzxX17/33ntZccUVM2XKlCTJd77znfz1r3+dq2uTZPvtt8+AAQO+8H6zZs2y5JJLZumll86yyy6bDTfcMBtvvHG+853vZLXVVpurtfv3759DDjkkSdKrV6/0799/tudOmjQpf/nLX3L77bfnhRdeyPvvv59x48Zl8cUXz/LLL5/VVlstG220Ubbccst861vfylJLLTVXM4wYMSI333xz7r///rz++uv54IMPUq1W0759+6yxxhrZeeeds++++2bFFVecq/X++3/X9957b7797W9/6TXDhg3LKquskiTZaqut8uSTT87VvRYGIiwAAAAAjeIvf/lLPvnkkyyzzDJzPPfqq6+er3tdc801tQCbJPfff3/efffddOzYcb7WnTp1aj799NN8+umneeeddzJ48OD069cvJ5xwQrbbbrv07t073bt3n697TPfss8+mZ8+eee21177w2dixYzN27Ni8+eabeeCBB5Ik7du3z/vvv/+la44fPz69e/fOpZdemokTJ37h82HDhmXYsGF58MEHc/rpp+eYY47JOeeck1atWjVo9p/97GfZeeedvxBnvy5EWAAAAAAWqBYtWmTy5MmZOHFirr/++hx33HFfev4nn3ySv/zlLzNd21D/HXGnTJmS/v3757TTTmvwWr/4xS+y3nrr1V6PHz8+o0ePztChQ/PMM89k4MCBmTJlSh577LEMGDAgxxxzTC6++OI0b968wfea7oUXXkiPHj0yduzYJEmHDh3yve99LxtssEGWWWaZfPbZZxkxYkReeOGFPPzwwxk9evRM0XlW/v3vf2f33XfP888/X3tv6623zi677JIVV1wxlUolw4cPz3333ZcnnngiEydOzG9/+9s88cQTueuuu7LCCivM9fxDhgzJn/70p/zgBz+Yt/8ACzkRFgAAAIAFavnll8/yyy9f2zU6pwh7ww03ZMKECUmSXXfdNXfccUeD7vf444/nX//6V5Jpj0C46667Mn78+PTr12+eIuzWW2+d7bfffrafv/322/nlL3+Zvn37plqt5tJLL83UqVPz+9//vsH3mu6II46oBdhevXrl8ssvT8uWLWd57uTJk/PQQw/l5ptvnu16EyZMyC677JIhQ4YkSVZZZZX0798/22677RfO/dnPfpYnn3wyvXr1ytChQ/P8889nl112ycCBA2c7w3QtW7bMpEmTMnXq1PTu3Tv77LNPFllkkbn81l8dfpgLAAAAgAXu0EMPTZIMHjw4L7300peeO30X66abbjrTDtS5ddVVV9WOjz/++Oy1115JkjfeeGOWz3mdXyuttFKuuOKK/PGPf6y9d9lll+XPf/7zPK33yiuvZPDgwUmSFVdcMVdeeeWXxs8WLVrk29/+9pc+wuHHP/7xTAH2iSeemGWAnW7rrbfO448/Xntm65AhQ/KTn/xkjrO3a9cuBx10UJJk6NChueKKBfcM4qZEhAUAAABggfvBD35QC4kzRtL/9tJLL9UC5PRw2xBjxozJLbfckiRZbbXVstVWW6VXr161z+f3WbNf5qCDDsqJJ55Ye3322Wdn6tSpDV7nn//8Z+14iy22mO+dpMOHD6/F0Eqlkv79+6dTp05zvK5Tp07p169f7bmul19+eUaMGDHH637+859nscUWSzLtUQ7Td/R+nYiwAAAAACxwyyyzTPbcc88kyfXXX59JkybN8rzpgbZly5bZf//9G3yfP/3pTxk/fnySpGfPnkmSHj16pHPnzkmSW265JWPGjGnwunPrtNNOq8Xmf/zjHxk0aFCD15jxGbhz+qGtufH73/8+n3/+eZJkp512+tIdsP9tu+22yw477JAk+fzzz3PppZfO8ZouXbrkmGOOSZKMGjUqF1544TxMvXATYQEAAABoFNN3tn700Ue58847v/D5pEmTcv311ydJ9t5777Rp06bB95gecSuVSi3CNmvWrPZP5MePH58bb7xxXsafK8stt1x22mmn2uvHHnuswWusvvrqteOnn346zz777HzNdP/999eOZ9wVPLcOPvjg2vGDDz44V9ecfvrpWWqppZIkF1xwQT744IMG33dhJsICAAAA0Ch69OiRLl26JJn1YwHuuOOOfPTRR0nm7VEEL7/8cp577rkkybbbbpuVV1659tmM8fHLHodQD1tssUXtePo8DbHxxhtn7bXXTjJt92mPHj3yk5/8JIMGDartaJ1bY8eOzd/+9rfa66222qrB82y55Za14yFDhmTcuHFzvKZdu3a1Z8j+5z//yTnnnNPg+y7MRFgAAAAAGkWzZs1yyCGHJEkeeOCBjBw5cqbPp4fZlVdeOd27d2/w+jPG1em7YKdba621stlmmyVJnn322fzjH/9o8Ppza6WVVqodz8sO0EqlkquvvjqtWrVKMi2knn/++dliiy3SunXrbLrppjnqqKPyxz/+8Qv/Df/be++9V3subcuWLWsRvCFWXnnl2jNep0yZkn//+99zdd2JJ56YFVZYIcm058kOGzaswfdeWImwAAAAADSaQw45JJVKJVOmTMkf//jH2vsjR47MAw88kGTaP3+f/mNQc2vSpEm57rrrkiStWrXKPvvs84VzFtRu2GWWWaZ2PH1nb0NtvvnmefbZZ/Otb31rpvcnTZqU559/PldccUV69eqVLl26ZMcdd8xTTz01y3U+/vjj2vG8PN5hVtfO7XdaYoklcuaZZ9bm7t279zzff2EjwgIAAADQaFZaaaXaLtd+/frV3u/fv3+mTp2aSqUy0zNI59aMjzLYa6+9suSSS37hnP3337+2o/Paa69t8D/tn1vVarV23NCYPKN11103jzzySP7+97/n7LPPzg477JCll156pnOmTp2ahx56KNtss02T/Cf/P/zhD2vPuL3hhhvy0ksvNfJEC4YICwAAAECjmv6813/961954oknkkyLsEnSvXv3mf45/9yacWfr7H58aplllsnuu++eJPnwww9n+eNg9fDJJ5/Ujtu1azff66277ro588wz8+CDD+aTTz7JG2+8keuuuy4/+MEPsuiiiyaZFn579+6dG264YaZr27ZtWzsePXr0PM8w47UN+U6LLLJILQ5PnTo1P/vZz+Z5hoWJCAsAAABAo9p7771r/7y9X79+GTBgQN54440k8/aDXMOHD8+DDz6YJOnUqVN69Ogx23MXxCMJZnz26XLLLVfXtSuVSlZbbbUceOCBuf766/PPf/4za665Zu3zPn36zHR+hw4d0qzZtCQ4YcKEvPPOOw2+57BhwzJx4sQkSfPmzWvPeZ1b++67b7p27Zokueeee/L44483eIaFjQgLAAAAQKNq2bJlDjjggCTJn//85/zf//1fkmnPHd17770bvF6/fv1qPz41cuTING/ePJVKZZZ/pu+ETWb942D1MHDgwNrx9B8DK2WVVVap7SJOkjfeeGOmCNy6detsuOGGtdeze3bsl5nx+2y00UZZYoklGnR9pVLJeeedV3t96qmnNniGhY0ICwAAAECjm77jdezYsbntttuSJAcccEBatmzZoHWq1epMz5ZtiClTpswUMOth1KhRtV25SbLddtvVdf1Z2XzzzdO6deva6/fee2+mz3feeefa8Yw/hja3ZvxvtNNOOzV8wCQ77rhjdthhhyTTou7tt98+T+ssLFo09gAAAAAAsMkmm2SDDTaY6YeaDjnkkAav8/DDD9d2fq622mr5n//5nzleM378+Jx//vlJpu2iPe200+brB7Rm9Mtf/rL2T/c32GCD4jthk2k7TVu0+H/Zb8YgmyRHH310LrjggkyePDn3339/nnzyyWy99dZztfYTTzxRi8qLLLJIjjnmmHme87zzzsumm26aarWa008/PXfdddc8r9XUibAAAAAANAk/+tGPctlllyWZ9izXTTfdtMFrzPhc12OPPTYnnXTSXF330EMP5cUXX8ybb76ZAQMGZPvtt2/wvf/btddem4svvrj2uk+fPvMUd0ePHp1WrVrVfnRrTgYMGFD74azFF188q6222kyfd+nSJUceeWQuvfTSVKvV9OrVK08++WQ6dOjwpeu+9957Ofjgg1OtVpMkRx11VDp37tzg7zNdt27dss8+++Tmm2/Oq6++WvddyE2JxxEAAAAA0CT06tUrgwYNyqBBg3Lrrbc2+PqPP/649s/aW7RokQMPPLBB957u6quvbvC9Z/TOO+/kqKOOSs+ePWvvHXfccfP0fNskGTRoUFZZZZWcf/75X3i0wH/729/+NtN9v/e976VVq1ZfOO+CCy6oPRt26NCh2WabbfLkk0/Odt2BAwdm2223zdChQ5NMexbs9N3D8+Occ86p7dr97W9/O9/rNVV2wgIAAADwlXD99dfX/un/zjvvnPbt28/1tT/4wQ9yyimnZPLkybnllltyySWXZKmllprluU8++WRtp2mSfPbZZ/n0008zdOjQDBo0KE8//XSmTJmSZNqjAY477rhcdNFF8/y9kuTdd9/NT37yk5x66qnZbLPNssUWW2TNNddM27ZtM3ny5AwfPjwDBgzI/fffX7t3586d87//+7+zXK9ly5a57777sttuu+WFF17Im2++mW222SbbbLNNdtlll6y44oqpVCoZPnx47r///gwYMKC2A7Zbt265++67s9hii83Xd0qSNdZYIz/84Q9z+eWXZ9y4cfO9XlMlwgIAAADwlTDjowhm3Nk6N5Zbbrnssssuueuuu/LZZ5/lT3/6U4488shZntu7d+85rlepVLLddtvlzDPPzLe+9a0GzfLf2rdvn44dO+bdd9/N1KlTM3DgwAwcOPBLr+nevXv69ev3pY8YWGGFFTJgwID07t07v//97zNx4sQ88cQTeeKJJ2Z5/mKLLZajjz4655xzTpZYYon5+k4zOvPMM/PHP/4x48ePr9uaTY0ICwAAAJBk6S2vaOwRmA8vvPBC/va3vyVJ2rRpkz322KPBa/Tq1av241BXXXXVbCPsjJo1a5bWrVtnqaWWynLLLZcNNtgg3bp1y3e+850vPIt1XnXt2jUjRozIc889l8ceeyyDBg3Ka6+9lpEjR2bs2LFZZJFFsvTSS2eNNdbIJptsku9///vZaqut5mrtJZZYIhdeeGFOOumk3Hzzzbn//vvz+uuv54MPPki1Ws1yyy2XNddcMzvvvHP23XffdOnSpS7faUYdOnTIiSeemF/+8pd1X7upqFSn7yPma6dz584ZOXJkOnXqlBEjRjT2ODBbfV8e09gjMAtHrD/rf5YDAAAAzMwPcwEAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCNtB//vOf3HrrrTnuuOOy5ZZbZrnllssiiyySpZZaKt/4xjfSs2fP3HfffalWq3O95htvvJEf//jHWW+99bL00kundevWWWuttXLsscdmyJAh5b4MAAAAAFBcpdqQWvg1d+GFF+b000/PhAkT5njuNttsk+uuuy5dunT50vP69u2bE088MZ999tksP2/evHnOPPPMnHnmmfM085fp3LlzRo4cmU6dOmXEiBF1Xx/qpe/LYxp7BGbhiPWXauwRAAAAYKHQorEHWJi8/vrrtQDbqVOn7LDDDunWrVvat2+fCRMmZNCgQbnuuusyduzYPPHEE9l+++0zaNCgtG/ffpbrXXfddTnyyCOTJM2aNcv++++fHj16pEWLFnnqqadyzTXXZOLEienTp08WW2yx/PSnP11g3xUAAAAAqA87YRvg6KOPztChQ3PKKaekR48eadbsi09zePvtt7PzzjvntddeS5Iccsghufrqq79w3gcffJDVV189Y8aMSbNmzXL77bdnjz32mOmcQYMGpUePHhk/fnxatGiRv//971lrrbXq9n3shGVhYSds02QnLAAAAMwdz4RtgHPPPTf3339/dtxxx1kG2CRZaaWVctNNN9Ve33TTTRk/fvwXzrvgggsyZsy0sHTsscd+IcAmyeabb55f/OIXSZLJkyfn7LPPrsfXAAAAAAAWIBG2Adq2bTtX52244Ya1Havjx4/PG2+88YVzZgy1J5100mzXOvzww7PEEkskSe68887ZPjsWAAAAAGiaRNhCllrq//0z3f8Op6+88krefvvtJMnaa6+dVVZZZbbrLLnkktlmm22SJOPGjcuAAQMKTAsAAAAAlCLCFjBp0qS8/vrrtdcrrbTSTJ+//PLLteNNN910juvNeM6M1wIAAAAATZ8IW8ANN9yQTz/9NEnStWvXrLDCCjN9Pv1Hu5J86S7YWZ0z47UAAAAAQNPXorEH+Kr54IMP8tOf/rT2+owzzvjCOaNHj64dL7vssnNcs127drO8tl4+//zzDB48eJafdejQIR06dKj7PQEAAADg60KEraNJkyble9/7XkaNGpUk2XPPPbPXXnt94byxY8fWjlu2bDnHdRdffPHa8X/+8586TDqzUaNGpVu3brP8rE+fPjnrrLPqfk8AAAAA+LoQYetk6tSpOfTQQ/PEE08kSVZbbbVcffXVjTzV3Gnfvn3uvffeWX5mFywAAAAAzB8Rtg6q1WqOOuqoXH/99UmSLl265KGHHsoyyywzy/Nbt25dO54wYcIc1//ss89qx0suueR8TvtFiyyySLp27Vr3dQEAAAAAP8w136rVao455phceeWVSZLOnTvnkUceycorrzzba9q0aVM7/vDDD+d4j48++miW1wIAAAAATZ8IOx+q1WqOPfbYXH755UmSTp065dFHH81qq632pdettdZateO33nprjveZ8ZwZrwUAAAAAmj4Rdh5ND7CXXXZZkqRjx4559NFHs/rqq8/x2vXXX792/Nxzz83x/BnPWW+99eZhWgAAAACgsYiw8+C/A2yHDh3y6KOPZo011pir69dZZ5106dIlSfLqq69m2LBhsz137NixtR/7atWqVbbbbrv5Gx4AAAAAWKBE2Hlw3HHH1QLsCiuskEcffTRrrrlmg9bYb7/9ascXXnjhbM/r27dvxo0blyTZY4890qpVq3mYGAAAAABoLCJsAx1//PH5/e9/n2RagH3sscfm6Tmtp5xySpZccskkyaWXXpo777zzC+c888wz6d27d5KkRYsW6dOnz3xMDgAAAAA0hhaNPcDC5Iwzzsgll1ySJKlUKjnhhBPy6quv5tVXX/3S67p27Vp7/MB07du3z+9+97scfPDBmTp1avbaa6/sv//+2XHHHdO8efM89dRTueaaazJhwoQkydlnn51vfOMbZb4YAAAAAFBMpVqtVht7iIXF9ttvnwEDBjT4un79+uXggw+e5WeXXXZZfvSjH9Vi639r3rx5Tj/99Jx99tkNvu+cdO7cOSNHjkynTp0yYsSIuq8P9dL35TGNPQKzcMT6SzX2CAAAALBQsBO2kR199NHZYYcdcvnll+e+++7L8OHDM3Xq1HTs2DE9evTIEUcckY033rixxwQAAAAA5pGdsF9jdsKysLATtmmyExYAAADmjh/mAgAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIQFAAAAAChIhAUAAAAAKEiEBQAAAAAoSIRtoClTpuTvf/97+vfvn+OPPz5bbLFFWrVqlUqlkkqlkoMPPniu1unfv3/tmrn5c9ZZZxX9XgAAAABAGS0ae4CFzb777pvbbrutsccAAAAAABYSImwDTZkyZabXbdu2Tbt27fKvf/1rntc8/vjj07179y895xvf+MY8rw8AAAAANB4RtoG++c1vZu211063bt3SrVu3rLLKKunfv38OOeSQeV6za9eu2XPPPes3JAAAAADQZIiwDXTaaac19ggAAAAAwELED3MBAAAAABQkwgIAAAAAFCTCNgG///3vs/baa6d169Zp1apVunTpkj322COXXXZZxo8f39jjAQAAAADzwTNhm4DnnntuptfDhw/P8OHDc9ddd6VPnz65+uqrs9tuuxW7/+eff57BgwfP8rMOHTqkQ4cOxe4NAAAAAF91Imwjat68ebbYYotss802WXPNNdO6deuMHj06L7zwQm6++eZ8/PHH+eCDD7LHHnvk+uuvzwEHHFBkjlGjRqVbt26z/KxPnz4566yzitwXAAAAAL4ORNhGsvXWW2fYsGHp3LnzFz774Q9/mP/93//N4YcfnptuuinVajWHHnpottpqq3Tp0qXus7Rv3z733nvvLD+zCxYAAAAA5o8I20hWX331L/18ySWXzPXXX5/3338/jz32WCZMmJBf//rXufTSS+s+yyKLLJKuXbvWfV0AAAAAwA9zNWnNmzfPOeecU3t99913N+I0AAAAAMC8EGGbuC222CItW7ZMkrzzzjsZP358I08EAAAAADSECNvENWvWLG3btq29Hj16dOMNAwAAAAA0mAjbxE2dOjWffPJJ7XWbNm0abxgAAAAAoMFE2CZu0KBB+eyzz5IknTt3TqtWrRp5IgAAAACgIUTYJmzq1Kk588wza6932223RpwGAAAAAJgXImwjGDhwYPr27ZsJEybM9pxx48alZ8+eefjhh5Mkiy22WH76058uqBEBAAAAgDpp0dgDLGzeeuutXHXVVTO999JLL9WOX3zxxZxxxhkzfd69e/d079699vr999/PkUcemZNPPjk77rhjunXrlhVXXDFLLLFEPv300wwePDg33nhjPvrooyRJpVLJH/7wh6y88srlvhgAAAAAUIQI20Bvv/12zj333Nl+/tJLL80UZZOkRYsWM0XY6caOHZvbb789t99++2zXW2GFFfKHP/whu+6667wPDQAAAAA0GhG2Eeywww6544478swzz+TZZ5/N8OHD89FHH2X06NFp1apV2rdvn65du2bXXXfNvvvum5YtWzb2yAAAAADAPKpUq9VqYw9B4+jcuXNGjhyZTp06ZcSIEY09DsxW35fHNPYIzMIR6y/V2CMAAADAQsEPcwEAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFNSiXgv17ds3Bx54YJZYYol6LQnAV0jfl8c09gjMwhHrL9XYIwAAAHzl1W0n7FFHHZWOHTvm6KOPzosvvlivZQEAAAAAFmp1fRzB2LFj07dv32yyySb55je/mauvvjrjx4+v5y0AAAAAABYqdYuwffr0SadOnVKtVlOtVvPCCy/k8MMPT8eOHXPcccflpZdeqtetAAAAAAAWGnWNsMOGDcudd96Z3XbbLc2aNUu1Ws2YMWNy2WWXZeONN84WW2yRa665JhMmTKjXbQEAAAAAmrS6Po6gWbNm2W233XLnnXfmrbfeyplnnpnOnTvXdsc+++yzOfTQQ9OxY8eccMIJ+cc//lHP2wMAAAAANDl1jbAz6ty5c84666wMGzYsd9xxR3bdddfa7tjRo0fnkksuyQYbbJCtt9461113XSZOnFhqFAAAAACARlMswtZu0KxZdt9999x1111566230rt375l2xw4cODC9evVKx44d86Mf/SivvfZa6ZEAAAAAABaY4hF2Rp07d87ZZ5+doUOH5rjjjqu9X61W88knn+Tiiy/OOuusk1133TUvvPDCghwNAAAAAKCIBRphP/jgg/z617/O2muvnUsvvTSVSiXVajVJsvjii9d2x953333ZbLPNcvrppy/I8QAAAAAA6m6BRNiHHnoo++67b1ZcccWcdtppefPNN1OtVtOiRYvsu+++efTRRzNmzJjcdttt2WmnnVKtVjN16tScd955+dOf/rQgRgQAAAAAKKJYhB01alTOO++8rL766tl5551z6623ZtKkSalWq1lxxRXzi1/8Iu+8805uvPHGbLfddmnevHn23HPP3HfffXnsscfSrl27VKvVXHzxxaVGBAAAAAAorkW9F3zwwQfTt2/f3HnnnZk8eXKSac98rVQq2XnnnXPMMcdk1113TbNms++/2267bX784x/n1FNP9UNdAAAAAMBCrW4R9le/+lX+8Ic/ZNiwYUlSe9Zru3btcuihh+bII4/MqquuOtfrrbvuukmSMWPG1GtEAAAAAIAFrm4R9vTTT5/ph7a22GKLHH300dlnn32y2GKLNXywFnXfpAsAAAAAsMDVtXS2atUqBx54YI4++uhsuOGG87XWdtttl7feeqtOkwEAAAAANI66RdhLLrkkBx10UJZccsm6rNeyZcustNJKdVkLAAAAAKCx1C3CHnPMMfVaCgAAAADgK6NZYw8AAAAAAPBVVredsOPHj0/v3r1TrVbz/e9/P1tuueUcr3n66adzyy23pHnz5jn33HOz6KKL1mscAAAAAIAmoW4R9pZbbslvf/vbLLroovnZz342V9esueaaueyyyzJp0qR069Yt+++/f73GAQAAAABoEur2OIJ77703SdKjR48st9xyc3XNsssumx122CHVajV//etf6zUKAAAAAECTUbcIO3jw4FQqlWy99dYNum76+S+88EK9RgEAAAAAaDLqFmFHjhyZJFl55ZUbdF2XLl2SJCNGjKjXKAAAAAAATUbdIuzkyZOTJM2bN2/YAM2mjTBx4sR6jQIAAAAA0GTULcK2a9cuSfLOO+806Lrhw4cnSdq0aVOvUQAAAAAAmoy6Rdi11lprnn5ga/r5a6yxRr1GAQAAAABoMuoWYXfcccckyeOPP5577rlnrq65++67M2DAgFQqley00071GgUAAAAAoMmoW4T94Q9/mFatWiVJDjjggNx2221fev6tt96aAw88MEnSsmXLHHHEEfUaBQAAAACgyWhRr4WWW265nHvuuTnppJMyduzY7LPPPunatWu++93vZp111knr1q0zduzYvPLKK7njjjsyePDgVKvVVCqV/PznP88KK6xQr1EAAAAAAJqMukXYJDnhhBMyfPjwXHjhhUmSwYMHZ/DgwbM8t1qtJklOPvnknHzyyfUcAwAAAACgyajb4wimu+CCC3LLLbdknXXWSbVane2fddddN7fffnv+93//t94jAAAAAAA0GXXdCTvd3nvvnb333jsvvPBCnnjiiYwYMSJjxozJUkstlc6dO2fbbbdN165dS9waAAAAAKBJKRJhp+vWrVu6detW8hYAAAAAAE1a3R9HAAAAAADA/yPCAgAAAAAUVPRxBJMmTcro0aMzYcKEuTq/S5cuJccBAAAAAFjg6h5h33rrrVx88cW599578+abb6Zarc7VdZVKJZMnT673OAAAAAAAjaquEfamm27KYYcdls8++yxJ5jrAAgAAAAB8VdUtwr788ss56KCDMmXKlFSr1bRs2TKbbLJJOnfunMUWW6xetwEAAAAAWKjULcKef/75mTx5ciqVSn74wx/m/PPPz9JLL12v5QEAAAAAFkp1i7ADBgxIpVLJ9ttvn759+9ZrWQAAAACAhVqzei00atSoJMl+++1XryUBAAAAABZ6dYuw0x890LZt23otCQAAAACw0KtbhF1nnXWSJMOHD6/XkgAAAAAAC726RdiePXumWq3mtttuq9eSAAAAAAALvbpF2F69emXbbbfN008/nd/97nf1WhYAAAAAYKFWtwhbqVRy2223Zdttt82JJ56YQw89NC+//HK9lgcAAAAAWCi1qNdCq666apJk8uTJqVarueaaa3LNNddkiSWWSNu2bdOs2Zf33kqlkjfffLNe4wAAAAAANAl1i7DDhg1LpVJJMi2oVqvVJMnYsWMzduzYOV4//VoAAAAAgK+SukXYLl26CKkAAAAAAP+lrjthAQAAAACYWd1+mAsAAAAAgC8SYQEAAAAAChJhAQAAAAAKqtszYf/bc889l/vvvz+vvPJKPv7443z++ed5+OGHZzrnww8/zKRJk9KyZcu0bdu21CgAAAAAAI2m7hH2jTfeyKGHHpqnnnqq9l61Wk2lUvnCub/61a9y0UUXZbnllsvIkSPTvHnzeo8DAAAAANCo6vo4gsGDB2eTTTbJU089lWq1WvszO0cffXSq1Wo++OCDPPDAA/UcBQAAAACgSahbhP3ss8+y5557ZsyYMWnevHlOO+20vPbaa7n55ptne83qq6+ejTbaKEny4IMP1msUAAAAAIAmo24R9sorr8yIESNSqVRy00035Zxzzskaa6yRRRZZ5Euv22abbVKtVvP888/XaxQAAAAAgCajbhH2jjvuSKVSyS677JK99tprrq9be+21k0x7liwAAAAAwFdN3SLsP/7xjyTJrrvu2qDr2rZtmyQZPXp0vUYBAAAAAGgy6hZhP/nkkyRJ+/btG3Tdl/1wFwAAAADAwq5uEXbppZdOkowZM6ZB140YMSJJ0q5du3qNAgAAAADQZNQtwq688spJkhdeeKFB1z388MNJknXWWadeowAAAAAANBl1i7A9evRItVrNTTfdNNe7YYcMGZL7778/lUolO+ywQ71GAQAAAABoMuoWYQ8//PC0aNEiH3/8cXr16pXJkyd/6flDhw7N97///VSr1bRq1SqHHnpovUYBAAAAAGgy6hZhV1111ZxyyimpVqu58847s9FGG+UPf/hDhg4dWjvnlVdeyX333ZcTTjghG264YYYOHZpKpZI+ffp4JiwAAAAA8JXUop6LnXvuuRk+fHiuv/76vPrqqznyyCOTJJVKJUmy/vrr186tVqtJkkMPPTSnnHJKPccAAAAAAGgy6rYTNpkWW6+99tpcdtllWWGFFVKtVmf7Z7nllsull16aK6+8sp4jAAAAAAA0KXXdCTvdkUcemUMOOSQPPPBAHn/88QwbNiyjR49O69at07lz52y33XbZZZdd0qpVqxK3BwAAAABoMopE2CRZdNFFs9tuu2W33XYrdQsAAAAAgCavro8jAAAAAABgZiIsAAAAAEBBIiwAAAAAQEF1eybsqquuOl/XVyqVvPnmm3WaBgAAAACgaahbhB02bFgqlUqq1eqXnlepVJLkC+dNfx8AaLo+ffrIxh6BWVh6yysaewQAAOBL1C3CdunSZY4hdcqUKfn4448zfvz4JNPCa8eOHdOiRd3GAAAAAABoUuq6E3ZuDRkyJL/5zW9y/fXXZ4011shtt92WNm3a1GsUAAAAAIAmo1F+mGujjTbKtddemwsvvDCPPfZY9t577zk+xgAAAAAAYGHUKBF2uhNPPDGbb755BgwYkGuuuaYxRwEAAAAAKKJRI2yS7LPPPqlWqyIsAAAAAPCV1OgRdsUVV0ySvPLKK408CQAAAABA/TV6hP3kk0+SJGPGjGnkSQAAAAAA6q/RI+yf//znJMnyyy/fyJMAAAAAANRfo0XYsWPH5vjjj89DDz2USqWS7bbbrrFGAQAAAAAopkW9Fjr00EPn6rxJkyZl5MiRefbZZzNhwoQkSfPmzXPKKafUaxQAAAAAgCajbhG2f//+qVQqc31+tVpNkrRs2TJXXnll1l9//XqNAgAAAADQZNQtwib/L6zOjVVXXTW77LJLTjjhhKy++ur1HAMAAAAAoMmoW4R966235uq8xRZbLG3atEnLli3rdWsAAAAAgCarbhF2pZVWqtdSAAAAAABfGc0aewAAAAAAgK8yERYAAAAAoCARFgAAAACgoLo9E/bnP/95vZb6gjPPPLPY2gAAAAAAJdUtwp511lmpVCr1Wm4mIiwAAAAAsLCqW4RNkmq1WjuuVCozvf5vc/p8xvMAAAAAABZWdYuwjz76aJLkd7/7XW677bY0a9YsO+20U3r06JHVV189SyyxRMaNG5c33ngjDz/8cB544IFMnTo1e++9d4477rh6jQEAAAAA0KTULcJut912Oemkk3L77bdn7bXXzo033pj1119/luf+6Ec/yt///vfst99+ue2229KlS5f85je/qdcoAAAAAABNRrN6LfTggw/m4osvTtu2bfPII4/MNsBOt9566+WRRx7JMsssk4suuigPPfRQvUYBAAAAAGgy6hZhL7/88lQqlRx22GFZfvnl5+qa5ZdfPocddliq1WquuOKKeo0CAAAAANBk1C3CPv/880mSjTbaqEHXbbzxxkmSZ599tl6jAAAAAAA0GXWLsKNGjUqSTJw4sUHXTT9/+vUAAAAAAF8ldYuwyyyzTJJkwIABDbpu+vlt2rSp1ygAAAAAAE1G3SLs5ptvnmq1muuuuy4DBw6cq2sGDRqU6667LpVKJZtvvnm9RgEAAAAAaDLqFmGPPPLIJMmUKVOy88475/LLL8/nn38+y3M///zzXHHFFfn2t7+dyZMnJ0mOPvroeo0CAAAAANBktKjXQjvvvHMOO+ywXHXVVRk3blyOPfbYnHbaadlqq62y+uqrp1WrVhk/fnzeeOONPPXUU/n0009TrVaTJIcddlh22mmneo0CAAAAANBk1C3CJknfvn3TqlWrXHLJJalWqxk9enTuueeeL5w3Pb5WKpUcf/zx+e1vf1vPMQAAAAAAmoy6PY4gmRZVL7744jz++OPZc889s+iii6ZarX7hz2KLLZa99torTzzxRC666KJUKpV6jgEAAAAA0GTUdSfsdFtttVW22mqrTJo0KX/729/y7rvvZuzYsWndunU6deqUDTbYIIsuumiJWwMAAAAANClFIux0iy66aDbddNOStwAAAAAAaNLq+jgCAAAAAABmVnQn7IgRI/LKK6/k448/zqRJk9KzZ8+StwMAAAAAaHKKRNirr746v/nNb/LPf/5zpvf/O8Kee+65GTBgQFZcccVcddVVJUYBAAAAAGhUdX0cwWeffZZdd901hx9+eP75z3+mWq3W/szKJptskoceeij9+/fPq6++Ws9RAAAAAACahLpG2J49e+bee+9NtVrNSiutlJ/97Gc56qijZnv+jjvumOWWWy5Jcvfdd9dzFAAAAACAJqFuEfbhhx/OrbfemkqlkgMOOCCvvfZazj333Oy8886zv3mzZtlxxx1TrVbz5JNP1msUAAAAAIAmo24Rtn///kmSVVddNf37988iiywyV9dtuOGGSeJxBAAAAADAV1LdIuxTTz2VSqWSnj17znWATZKOHTsmSf7973/XaxQAAAAAgCajbhH2/fffT5KstdZaDbquZcuWSZIJEybUaxQAAAAAgCajbhG2efPmSZKpU6c26LqPP/44SdKmTZt6jQIAAAAA0GTULcIuv/zySZI33nijQde98MILSZIVV1yxXqMAAAAAADQZdYuwW265ZarVav7yl7/M9TXjxo3Ln//851QqlWy99db1GgUAAAAAoMmoW4TdZ599kiQvvvhirr766rm65uijj84nn3ySJDnwwAPrNQoAAAAAQJNRtwi72267ZfPNN0+1Ws1RRx2VX/3qVxk7duwsz33xxRez66675vrrr0+lUskuu+ySb37zm/UaBQAAAACgyWhRz8VuuummbLbZZvn3v/+dM844I7/4xS9qz4pNkk033TQjRozIqFGjkiTVajVdunRJ//796zkGAAAAAECTUbedsMm0H9d65plnajtiJ0yYkHfeeSeVSiVJMnjw4Lz//vupVqupVqvZbLPN8vTTT2fZZZet5xgAAAAAAE1GXSNsMi3EPv3007njjjuy9957p127drXoWq1W07p16+y66665+eabM3DgwHTs2LHeIwAAAAAANBl1fRzBjHbffffsvvvuSZLx48dn9OjRad26dZZaaqlStwQAAAAAaHLqFmG7d++eJNl6663z85//fKbPWrVqlVatWtXrVgAAAAAAC426RdgBAwYkSfbee+96LQkAAAAAsNCr2zNhp/+41gorrFCvJQEAAAAAFnp1i7CrrrpqkuT999+v15IAAAAAAAu9ukXYPffcM9VqNXfffXe9lgQAAAAAWOjVLcIeddRRWXHFFfPAAw/kxhtvrNeyAAAAAAALtbpF2KWXXjp33HFHOnfunJ49e+bkk0/OsGHD6rU8AAAAAMBCqUW9FurevXuSaTF2+PDhueiii3LRRRelY8eO6dy5cxZffPEvvb5SqeThhx+u1zgAAAAAAE1C3SLsY489lkqlkiS1/1mtVvPuu+/m3Xff/dJrq9Vq7RoAAAAAgK+SukXYZFpMnZv3AAAAAAC+LuoWYadOnVqvpQAAAAAAvjLq9sNcAAAAAAB80TxF2O7du6dHjx55+umn6z0PAAAAAMBXyjxF2MceeyyPPfZYPvzww9me07Vr13Tr1i0DBgyY5+EAAAAAABZ2xR5HMGTIkAwZMiSffvppqVs0iilTpuTvf/97+vfvn+OPPz5bbLFFWrVqlUqlkkqlkoMPPrjBa77xxhv58Y9/nPXWWy9LL710WrdunbXWWivHHntshgwZUvfvAAAAAAAsOHX7Ya6vi3333Te33XZb3dbr27dvTjzxxHz22Wczvf/666/n9ddfzxVXXJEzzzwzZ555Zt3uCQAAAAAsOCJsA02ZMmWm123btk27du3yr3/9q8FrXXfddTnyyCOTJM2aNcv++++fHj16pEWLFnnqqadyzTXXZOLEienTp08WW2yx/PSnP63LdwAAAAAAFhwRtoG++c1vZu211063bt3SrVu3rLLKKunfv38OOeSQBq3zwQcf5Nhjj00yLcDefvvt2WOPPWqf9+zZM4ccckh69OiR8ePH54wzzsiee+6ZtdZaq67fBwAAAAAoS4RtoNNOO60u61xwwQUZM2ZMkuTYY4+dKcBOt/nmm+cXv/hFTj755EyePDlnn312brjhhrrcHwAAAABYMIr9MBdf7qabbqodn3TSSbM97/DDD88SSyyRJLnzzju/8OxYAAAAAKBpE2EbwSuvvJK33347SbL22mtnlVVWme25Sy65ZLbZZpskybhx4zJgwIAFMiMAAAAAUB/zFWErlUpdzvm6efnll2vHm2666RzPn/GcGa8FAAAAAJq++Xom7J577vmln1er1TmeM12lUsnkyZPnZ5yFxmuvvVY7/rJdsLM6Z8ZrAQAAAICmb75/mKtarc7y/Rl3wM7unK+r0aNH146XXXbZOZ7frl27WV5bL59//nkGDx48y886dOiQDh061P2eAAAAAPB1Mc8Rdk5hVXidvbFjx9aOW7ZsOcfzF1988drxf/7zn7rPM2rUqHTr1m2Wn/Xp0ydnnXVW3e8JAAAAAF8X8xRhp06dWu85aETt27fPvffeO8vP7IIFAAAAgPkz348joOFat25dO54wYcIcz//ss89qx0suuWTd51lkkUXStWvXuq8LAAAAACTNGnuAr6M2bdrUjj/88MM5nv/RRx/N8loAAAAAoOkTYRvBWmutVTt+66235nj+jOfMeC0AAAAA0PSJsI1g/fXXrx0/99xzczx/xnPWW2+9IjMBAAAAAGWIsI1gnXXWSZcuXZIkr776aoYNGzbbc8eOHZsnnngiSdKqVatst912C2JEAAAAAKBORNhGst9++9WOL7zwwtme17dv34wbNy5Jsscee6RVq1bFZwMAAAAA6keEbSSnnHJKllxyySTJpZdemjvvvPML5zzzzDPp3bt3kqRFixbp06fPAp0RAAAAAJh/LRp7gIXNW2+9lauuumqm91566aXa8Ysvvpgzzjhjps+7d++e7t27z/Re+/bt87vf/S4HH3xwpk6dmr322iv7779/dtxxxzRv3jxPPfVUrrnmmkyYMCFJcvbZZ+cb3/hGoW8FAAAAAJQiwjbQ22+/nXPPPXe2n7/00kszRdlk2i7W/46wSdKrV6+MHz8+P/rRjzJhwoTccMMNueGGG2Y6p3nz5jn99NNz2mmn1ecLAAAAAAALlAjbyI4++ujssMMOufzyy3Pfffdl+PDhmTp1ajp27JgePXrkiCOOyMYbb9zYYwIAAAAA80iEbaDtt98+1Wq1rmuuscYa+c1vfpPf/OY3dV0XAAAAAGh8fpgLAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKCgFo09AABfDxu8fUJjj8CsrN+vsSegEXz69JGNPQKzsPSWVzT2CAAAFGInLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQSIsAAAAAEBBIiwAAAAAQEEiLAAAAABAQS0aewAoqe/LYxp7BGbjiPWXauwRAAAAABYIO2EBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAApq0dgDAAALj1c/ntTYIzALmzf2AAAAwJeyExYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgERYAAAAAoCARFgAAAACgIBEWAAAAAKAgEbYRbb/99qlUKnP9Z9iwYY09MgAAAADQQCIsAAAAAEBBLRp7AKa5/fbb53hO+/btF8AkAAAAAEA9ibBNxJ577tnYIwAAAAAABXgcAQAAAABAQSIsAAAAAEBBIiwAAAAAQEEibBOx2267pVOnTll00UWzzDLLZN11183hhx+eRx99tLFHAwAAAADmgwjbRPz1r3/Nu+++m88//zyjR4/OK6+8kj/84Q/p3r17evTokffee6+xRwQAAAAA5kGLxh7g626ZZZbJjjvumE022SSdOnVK8+bNM3LkyDz88MO59957U61W88gjj2SLLbbIoEGDssIKK9R9hs8//zyDBw+e5WcdOnRIhw4d6n5PAAAAAPi6EGEb0a9+9at069Ytiy666Bc++9GPfpTnn38+3/ve9/LOO+/k7bffzqGHHpp77rmn7nOMGjUq3bp1m+Vnffr0yVlnnVX3ewIAAADA14UI24i22GKLL/18k002yX333ZeNN944EydOzL333pvnnnsum266aV3naN++fe69995ZfmYXLAAAAADMHxG2iVt77bVz0EEH5Q9/+EOS5O677657hF1kkUXStWvXuq4JAAAAAEzjh7kWAt/61rdqx6+++mojTgIAAAAANJQIuxBYbrnlasejR49uvEEAAAAAgAYTYRcCH374Ye24TZs2jTcIAAAAANBgIuxC4NFHH60dr7XWWo04CQAAAADQUCJsE/f666/n2muvrb3ebbfdGnEaAAAAAKChRNhG8n//9395+umnv/ScF198MTvvvHMmTJiQJNlpp52y2WabLYjxAAAAAIA6adHYA3xdPfLIIznhhBOy2mqrZYcddsh6662Xdu3apXnz5nn33Xfz8MMP55577snUqVOTJCuttFL69evXyFMDAAAAAA0lwjayN998M2+++eaXnrPzzjvn6quvTseOHRfQVAAAAABAvYiwjeQ3v/lNdt999zzzzDP529/+llGjRuXDDz/MxIkTs/TSS2fllVfOFltskQMPPNAjCAAAAABgISbCNpLVVlstq622Wg477LDGHgUAAAAAKMgPcwEAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFCTCAgAAAAAUJMICAAAAABQkwgIAAAAAFNSisQcAAAAa5tOnj2zsEZiNpbe8orFHAACaIDthAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKEmEBAAAAAAoSYQEAAAAAChJhAQAAAAAKatHYAwAAAHwdffr0kY09ArOx9JZXNPYIAHzF2AkLAAAAAFCQCAsAAAAAUJAICwAAAABQkAgLAAAAAFCQCAsAAAAAUJAICwAAAABQkAgLAAAAAFCQCAsAAAAAUJAICwAAAABQkAgLAAAAAFCQCAsAAAAAUJAICwAAAABQkAgLAAAAAFCQCAsAAAAAUJAICwAAAABQkAgLAAAAAFCQCAsAAAAAUJAICwAAAABQkAgLAAAAAFCQCAsAAAAAUJAICwAAAABQkAgLAAAAAFCQCAsAAAAAUJAICwAAAABQkAgLAAAAAFCQCAsAAAAAUJAICwAAAABQkAgLAAAAAFCQCAsAAAAAUJAICwAAAABQkAgLAAAAAFCQCAsAAAAAUJAICwAAAABQkAgLAAAAAFCQCAsAAAAAUJAICwAAAABQkAgLAAAAAFBQi8YeAGBONnj7hMYegVlZv19jTwAAAFDT9+UxjT0Cs3DE+ks19ghNgp2wAAAAAAAFibAAAAAAAAWJsAAAAAAABYmwAAAAAAAFibAAAAAAAAWJsAAAAAAABYmwAAAAAAAFibAAAAAAAAWJsAAAAAAABYmwAAAAAAAFibAAAAAAAAWJsAAAAAAABYmwAAAAAAAFibAAAAAAAAWJsAAAAAAABYmwAAAAAAAFibAAAAAAAAWJsAAAAAAABYmwAAAAAAAFibAAAAAAAAWJsAAAAAAABYmwAAAAAAAFibAAAAAAAAWJsAAAAAAABYmwAAAAAAAFibAAAAAAAAWJsAAAAAAABYmwAAAAAAAFibBNxJ133pl99tknK6+8clq2bJn27dtnyy23zPnnn58xY8Y09ngAAAAAwDwSYRvZ2LFj893vfjff/e53c8stt+Ttt9/OxIkT88EHH2TgwIH5yU9+kvXWWy+DBg1q7FGbrPfeey9nnXVW3nvvvcYeha+5Dz8enz/c8GI+/Hh8Y4/C15y/izQF//5wfH511fP594f+HtK4/F2kqfD/t9AU+HtIU/F1/LsowjaiKVOmZJ999smdd96Z/H/t3XlQVff5x/EPXCLI6gpXRcWSiqBGxdiKmkKNuKCSaGpim0atKTWtJjNNk2qt0ZBUjU1tG5u4pC5xi9FOGqlGiRsWJWJN0URHjaYFw1gDbhgwItv5/eHPMyA73IUL79cMM9/jec7Jc+Y+PuY+93KOpKCgIM2bN0/vvvuu3nzzTQ0dOlSSlJ2drbi4OJ05c8aZ6TZZly5dUmJiYov6i4um6eq1W1qz5VNdvXbL2amghaMW0RR8dfUbLVmboa+uMviCc1GLaCp434KmgDpEU9ESa9HD2Qm0ZKtXr1ZycrIkKSIiQgcOHFBQUJC5f+bMmXrhhRe0dOlSXb9+XTNmzFBqaqqz0gUAAAAAAADQAHwT1klKS0uVmJhobm/cuLHCAPauJUuWqH///pKkQ4cOac+ePY5KEQAAAAAAAIANMIR1ktTUVPMr19HR0YqMjKwyzmKx6LnnnjO3t2zZ4pD8AAAAAAAAANgGQ1gn2b17t7mOi4urMXbMmDFVHgcAAAAAAACg6WMI6yQnT54014MGDaox1mq1qmvXrpKknJwcXb582a65AQAAAAAAALAdN8MwDGcn0RJ961vfUmZmpiQpMzNTISEhNcZHR0ebD+U6dOiQhg0b1ugcWrVqpeLiYrm7u6tDhw5VxlgsFrm7N+1ZfXFxsXJzcxUYGKj77ruvwr6bxZR3U+Vzn1udY4sKr9sxE9spKSnT9RuFahvgJQ+Ppv33xhZaebWtV7yrvI7NQX1qkdexeajv62gU5dknkXKKSkp15XqhOrT1UisPi93/e82BW6s2dY51xGvYXDi6Fnkdm4f6vI51VdP7FsBRmnsdMgNomqp6/98ca9FqteqTTz6pdr+HA3NBOXl5eea6ugFoee3bt6/y2MYoLS2VJJWVlSk3N9cm53Sm5nANLUmesxOwo+s3Cp2dgoN84+wEUIu61SKvY/PQdF/HK9dbSk+0hZvOTqBZc1wt8jo2D/Z7HXnfgqaAOoQj5dWwryXVIkNYJykoKDDXXl5etca3bt3aXOfn59skBy8vLxUWFsrd3V3t2rWrMsYVvgkLAAAAAAAAOJPVaq1xP0PYFuzmTT6lBwAAAAAAAOyNrzg6ia+vr7kuLKz9V6Nu3bplrv38/OySEwAAAAAAAADbYwjrJG3atDHXV65cqTX+6tWrVR4LAAAAAAAAoGljCOskYWFh5jozM7PW+PIx5Y8FAAAAAAAA0LQxhHWSvn37mutjx47VGJuTk6Ps7GxJUmBgoDp27GjX3AAAAAAAAADYDkNYJxk9erS53r17d42xu3btMtdxcXF2ywkAAAAAAACA7TGEdZLo6GhZrVZJ0sGDB5WRkVFlXGlpqZYtW2ZuT5482SH5AQAAAAAAALANhrBOYrFYNH/+fHN7ypQpys3NrRQ3Z84cnThxQpI0dOhQjRo1ylEpAgAAAAAAALABN8MwDGcn0VKVlJQoLi5Oe/fulSRZrVYlJCQoIiJC165d05YtW3T48GFJUps2bXT48GH17t3bmSkDAAAAAAAAqCeGsE6Wn5+vH/3oR9q5c2e1McHBwdq6dauGDBniwMwAAAAAAAAA2AJD2CYiKSlJGzZs0LFjx5Sbmys/Pz+FhoZq4sSJmjFjhgICApydIgAAAAAAAIAGYAgLAAAAAAAAAHbEg7kAAAAAAAAAwI4YwgIAAAAAAACAHTGEBQAAAAAAAAA7YggLAAAAAAAAAHbEEBYAAAAAAAAA7IghLFzWP/7xD02aNEkhISHy8vJSYGCghgwZotdff11ff/21s9NDMxcTEyM3N7c6/2RlZTk7ZbiY0tJSnTp1Su+8846effZZRUVFydvb26ypadOm1fucX3zxhV588UX16dNHAQEB8vX1VVhYmGbOnKkTJ07Y/Brg+mxVh++88069eubLL79s1+uC68nPz9f777+vWbNmaciQIerYsaPuu+8++fv7q1evXpoyZYqSk5NlGEadz0lPREPYqhbpi2iMY8eO6a233tK0adM0aNAghYSEyNfXV56engoKClJMTIwSExN14cKFOp/z0qVLevnllzVw4EC1b99e3t7eCg0N1bRp05SammrHq4Ers1UtHjx4sF49sSHvhZoEA3Ax+fn5Rnx8vCGp2p+uXbsaR44ccXaqaMaio6NrrMF7fzIzM52dMlzMxIkTa6ypqVOn1ut8q1atMlq3bl3t+SwWi5GYmGifi4HLslUdrlu3rl49c8GCBXa9LriWpUuXGl5eXnWqnYceesi4cOFCreekJ6IhbFmL9EU0ho+PT53qxtPT01i0aFGt59u+fbvRtm3bGs81Y8YMo6SkxAFXB1diq1pMSUmpV0+s73uhpsJDgAspLS3VpEmTlJycLEkKCgpSQkKCIiIidO3aNW3ZskVpaWnKzs5WXFyc0tLSFB4e7uSs0dx98MEHtcYEBgY6IBM0J6WlpRW227Vrp/bt2+v8+fP1PtemTZs0Y8YMSZK7u7smT56shx9+WB4eHkpLS9P69et1+/ZtLViwQJ6enpo9e7ZNrgGuz5Z1eNezzz6r4cOH1xjTq1evBp8fzc+5c+dUWFgoSerSpYtGjBihgQMHKjAwUIWFhUpPT9emTZtUUFCgQ4cOKSYmRunp6dX+20tPREPZuhbvoi+iIQIDA/Wd73xH/fr1U48ePRQQEKDi4mJlZWXpww8/VFpamm7fvq25c+equLhY8+fPr/I8KSkpevzxx1VUVCRJGjt2rOLj4+Xj46OMjAytWbNGN27c0KpVq+Tm5qYVK1Y48jLhAmxVi3c98cQTmjx5co0x3bp1s+UlOI6zp8BAfaxcudL85CMiIsL46quvKsX86le/qvAJNGAP5b8JC9jDwoULjTlz5hh/+9vfjP/+97+GYVT81kxdP/3Nzc01/P39DUmGu7u7kZSUVCnmyJEjhre3tyHJ8PDwMM6ePWvLS4ELs1Udlj9m3bp19ksYzdIzzzxjjBw50tizZ49RWlpaZUxWVpYRFhZm1tlPfvKTKuPoiWgMW9YifRGNcfLkSaOsrKzGmPXr1xtubm5mL7t48WKlmMLCQiMkJMSsxb/85S+VYj7//HPDarWaMfv377fZdcD12aoWy38Ttjl/8597wsJllJaWKjEx0dzeuHGjgoKCKsUtWbJE/fv3lyQdOnRIe/bscVSKAGAzc+fO1eLFi/WDH/xAPXr0aPB5/vCHP5j3yZ45c6bi4+MrxQwePFivvvqqJKmkpKRCr0XLZqs6BBpj4cKF+uijjxQbGyt396rfvnTv3l1bt241t7du3apvvvmmUhw9EY1hy1oEGqNPnz5yc3OrMWbKlCkaN26cpDu97O5vk5a3du1a89kV48eP16xZsyrF9OzZU2+99Za5PW/evEZkjubGVrXYUjCEhctITU3VpUuXJEnR0dGKjIysMs5isei5554zt7ds2eKQ/ACgKSr/RvCXv/xltXEJCQny8fGRdOfBh7du3bJ7bgBQF+3atatTXL9+/RQWFiZJ+uabb/TFF19UiqEnojFsWYuAI/Tu3dtcf/XVV5X2v/fee+b6+eefr/Y8jz76qEJCQiRJR44cqdcDvwCp9lpsKRjCwmXs3r3bXMfFxdUYO2bMmCqPA4CW5PTp0+b/JIeHh9f4TUY/Pz899NBDkqSbN2/qn//8p0NyBABb8vf3N9f3Dk7piXCkmmoRcJTyHwBYrdYK+/Lz83X48GFJFXteVdzd3TV69Ghzm/fYqK+aarElYQgLl3Hy5ElzPWjQoBpjrVarunbtKknKycnR5cuX7ZobWrZx48apS5cuatWqldq2bavevXsrISFBKSkpzk4NLVx9+ua9MeWPBWxp+fLlCg8Pl6+vr7y9vdWtWzfFx8drxYoV/MouGqWoqEjnzp0zt7t3715hPz0RjlJbLd6Lvgh72LFjh/kAYS8vL40dO7bC/tOnT6usrEySNGDAAFkslhrPR09EQ9VWi/d6//331b9/f/n7+8vLy0udO3fWyJEj9frrr+vatWuOSNluPJydAFBXn3/+ubmuy33pevTooezsbPPYjh072i03tGwffvihuc7Ly1NeXp5Onz6t1atXa/jw4dq0aZM6derkxAzRUjWkb1Z1LGBLx44dq7CdnZ2t7Oxs7dixQwsWLNDatWvN+4YB9fHuu+/qxo0bkqTIyMhK37ShJ8JRaqvFe9EX0RipqanmYKqoqEjZ2dnas2eP+WwUDw8PrVy5stLzVOiJsLWG1uK9Tp06VWH70qVLunTpkvbu3atXXnlFb7zxhqZPn26fi7AzhrBwGXl5eea6Q4cOtca3b9++ymMBW2nbtq1iY2P14IMPqkuXLrJYLLp48aL279+v3bt3yzAMHThwQFFRUUpPT2/Rv3YB56BvoimxWCyKiorSQw89pJ49e8rX11d5eXn697//rW3btunatWu6fPmy4uPjtXnzZv3whz90dspwIZcvX9bs2bPN7aoeHENPhCPUpRbvoi/CFn7961/r6NGjlf7czc1N0dHRSkxM1Pe+971K++mJsLWG1mL5uMjISMXExCg8PFwBAQEqKCjQZ599pm3btunixYsqKCjQ008/rdzcXM2ZM8eel2MXDGHhMgoKCsy1l5dXrfGtW7c21/n5+XbJCS3X4sWLNXDgQLVq1arSvueff16ffPKJHnvsMX355Ze6cOGCpk+frl27djkhU7Rk9E00FcOGDVNWVpaCg4Mr7fvpT3+q3//+90pISNDWrVtlGIamT5+uoUOHqlu3bk7IFq6mqKhIjz32mHJzcyXdeYDMhAkTKsXRE2Fvda1Fib4I++vSpYtiY2P17W9/u8r99EQ4Sm21KElhYWE6e/asevbsWeX+1157TbNnz9af//xnSdLcuXMVExOjwYMH2yNlu+GesADQAFFRUVUOYO968MEHlZycLE9PT0l3bl5/76+aAUBLcf/991c5aLjLz89PmzdvVkxMjCSpsLBQS5YscVB2cGVlZWWaPn26Dh06JEkKDQ3V2rVrnZwVWqL61iJ9EbaSnp4uwzBkGIYKCgp04sQJvfLKK8rPz9dvf/tb9e3bV/v27XN2mmgBGlOLnTp1qnYAK0mtWrXSn/70Jz311FOSJMMw9Oqrr9rlOuyJISxchq+vr7kuLCysNb78U0j9/PzskhNQk/DwcPMfCUnauXOnE7NBS0TfhCuxWCz63e9+Z27TM1EbwzD0zDPPaPPmzZKkbt26ad++fWrbtm2V8fRE2Et9a7Gu6IuoLx8fH/Xr108vvfSSjh8/rs6dO+vq1asaO3ZspYdp0RNhT/WpxfpYtGiR3NzcJEn79++vUJeugCEsXEabNm3M9ZUrV2qNv3r1apXHAo70/e9/31yfOXPGiZmgJaJvwtVERUWZvxL55Zdf8lRwVMswDP3iF7/QX//6V0lScHCwDhw4oJCQkGqPoSfCHhpSi/VBX0RD9ejRQ6+99pqkO7fKWLhwYYX99EQ4Sm21WB/BwcG6//77JUm3b99WZmamTXJ0FIawcBlhYWHmui5/0crHlD8WcKSOHTuaa25gD0ejb8LVuLu7q127duY2fRNVMQxDM2fO1MqVKyXduddcSkqKQkNDazyOnghba2gt1gd9EY0xZswYc33w4MEK++iJcKSaarG+XPk9NkNYuIy+ffua69rurZmTk6Ps7GxJUmBgYIW/pIAjlf9UmU+M4Wj16Zv3xvTp08cuOQE1KSsr0/Xr181t+ibudXfotWLFCklS586dlZKSYn4rpib0RNhSY2qxPuiLaIzytw0oX0eSFBERIXf3OyOh48ePq7S0tMZz0RPRGDXVYn258ntshrBwGaNHjzbXu3fvrjG2/FPo4+Li7JYTUJuUlBRzzSfGcLSIiAjzKcpnzpxRVlZWtbEFBQXmw0S8vb0VHR3tiBSBCtLT0817ewUHB8vb29vJGaEpuXfo1alTJ6WkpNT4tOXy6ImwlcbWYn3QF9EY58+fN9f3fjHJz89PQ4cOlSTl5+fr8OHD1Z6nrKxMH330kbld/luNQF3UVIv1cfHiRfNcnp6eNrv1i6MwhIXLiI6OltVqlXTn6+sZGRlVxpWWlmrZsmXm9uTJkx2SH3Cvc+fOaePGjeb2uHHjnJgNWqonnnjCXP/xj3+sNu7tt9/WzZs3JUnx8fG8yYPDlZWVaf78+eY2PRP3mjVrljn0slqtSklJqfFJylWhJ8IWbFGLdUFfRGPdvVWGJHPgWl7598pLly6t9jzbt283b0cwePBglxt8wflqq8W6eumll2QYhqQ7z19xuX+fDcCFLF++3JBkSDJ69+5t5OTkVIp54YUXzJihQ4c6IUs0d2+88YaRlpZWY0xGRoYREhJi1uLIkSMdlB2as3Xr1pk1NXXq1Dodk5OTY/j5+RmSDHd3dyMpKalSTHp6uuHt7W1IMjw8PIwzZ87YOHM0J/Wtw48//thYtWqVcevWrWpjCgoKjCeffNI8r6enp5GZmWm7pOHyZs2aZdaH1Wo1zp4926Dz0BPRWLaoRfoiGmPFihXGgQMHjLKysmpjSkpKjMWLFxtubm5mDR08eLBS3K1bt4xu3bqZMW+++WalmHPnzhlWq9WM2bt3r02vB67LVrV4/vx5Y8mSJcaNGzeqPU9RUZHx4osvmueQVOt78qbIzTD+f4QMuICSkhLFxcVp7969ku588pyQkKCIiAhdu3ZNW7ZsMX+Nok2bNjp8+LB69+7tzJTRDD366KNKSkpSaGioRowYoT59+qh9+/ayWCz63//+p/3792vXrl0qKyuTJHXv3l0ff/yxOnfu7OTM4UoyMzO1Zs2aCn/22WefaceOHZKkBx54QOPHj6+wf/jw4Ro+fHilc61fv17Tpk2TdOcBH5MnT1ZsbKwsFovS0tK0fv16FRYWSpIWLlyouXPn2uGK4IpsUYfbt2/XhAkT5Ovrq9jYWA0cOFBdu3aVj4+Pbty4oYyMDL333nvmU5fd3Ny0YcMG/fjHP7bz1cFVzJs3z3ySspubmxYtWqRevXrVelxkZKR5+4Hy6IloKFvVIn0RjTFt2jStX79eXbt2VWxsrPr27avAwEC1atVKeXl5OnXqlJKSkirccuU3v/mNFi1aVOX59u3bp7i4OBUXF0u6843r+Ph4+fj4KCMjQ6tXr9aNGzckSQkJCXr77bftfo1wDbaqxRMnTmjAgAHy9PTU8OHDNWjQIPXo0UN+fn4qKCjQyZMntW3bNvO5P5IL//vs7CkwUF9ff/21MW7cuAqfgNz7Exwc7JKfisA1PPLIIzXWX/mfUaNGGRcvXnR2ynBBKSkpda6zuz8LFiyo9nzLly83vLy8qj3WYrEY8+fPd9wFwiXYog4/+OCDOh9rtVqNnTt3Oudi0WRFR0fXuw4lGevWrav2nPRENIStapG+iMaYOnVqnesnICDAWL58ea3n/Pvf/260adOmxnMlJCQYJSUlDrhCuApb1eLx48frfB5/f39jzZo1Dr5S2/EQ4GL8/Py0Y8cOJSUlacOGDTp27Jhyc3Pl5+en0NBQTZw4UTNmzFBAQICzU0UztXTpUo0fP15Hjx7Vp59+qtzcXF25ckW3b99WQECAQkJCFBUVpSeffFLf/e53nZ0uIEn6+c9/rhEjRmjlypVKTk5Wdna2ysrK1LlzZz388MP62c9+pgEDBjg7TTRDI0aMUFJSko4ePap//etfys7O1tWrV5WXlydvb28FBgYqMjJSY8eO1eOPPy4vLy9np4wWgJ4IZ6IvojGWLVumRx55RKmpqTp+/Lj+85//6MqVKyouLpavr6+CgoL0wAMPaNSoUZo0aVKd3hdPmDBBgwcP1ooVK7Rjxw5lZWWpsLBQnTp10rBhw/T000/zgEJUYqtaDA8PV3Jyso4ePaqjR48qKytLV65c0fXr1+Xp6akOHTqoX79+io2N1VNPPSV/f38HX6ntcDsCAAAAAAAAALAjd2cnAAAAAAAAAADNGUNYAAAAAAAAALAjhrAAAAAAAAAAYEcMYQEAAAAAAADAjhjCAgAAAAAAAIAdMYQFAAAAAAAAADtiCAsAAAAAAAAAdsQQFgAAAAAAAADsiCEsAAAAAAAAANgRQ1gAAAAAAAAAsCOGsAAAAAAAAABgRwxhAQAAAAAAAMCOGMICAAAAAAAAgB0xhAUAAAAAAAAAO/o/IXLpoO0WzqAAAAAASUVORK5CYII=\n", "text/plain": [ "
" ] }, "metadata": { "filenames": { "image/png": "/Users/folgert/projects/hda/_build/jupyter_execute/intro-probability/notebook_24_0.png" } }, "output_type": "display_data" } ], "source": [ "df_known.groupby('AUTHOR')['by'].plot.hist(\n", " alpha=0.6, range=(0, 35), rwidth=0.9, legend=True);" ] }, { "cell_type": "markdown", "id": "f98536dc", "metadata": {}, "source": [ "\n", "\n", "Looking at the distribution of frequencies above, it is clear that there is\n", "evidence of a systematic difference in the rate the two writers use the word\n", "*by*. A high rate of *by* for Hamilton (the 75th percentile is 12) is less than\n", "the typical rate Madison uses the word (the median for Madison is 12.5).\n", "\n", "There is one case, Federalist No. 83, clearly visible in the histogram, which\n", "is written by Hamilton and features a very high rate of *by*. This high rate has\n", "a simple explanation. Federalist No. 83's topic is the institution of trial\n", "*by jury* and understandably features the two-word sequence *by jury* 36\n", "times. If we ignore the *by* instances which occur in the sequence *by jury*,\n", "*by* occurs about 7 times per 1,000 words, safely in the region of expected\n", "rates from Hamilton given the other texts. Federalist No. 83 provides an\n", "excellent cautionary example of an allegedly content-free \"function word\" *by*\n", "turning out to be intimately connected to the content of a text. We can verify\n", "that the outlier is indeed Federalist No. 83 by querying our ``DataFrame``:" ] }, { "cell_type": "code", "execution_count": 14, "id": "7683a041", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "83 34.0\n", "Name: by, dtype: float64\n" ] } ], "source": [ "print(df_known.loc[df_known['by'] > 30, 'by'])" ] }, { "cell_type": "markdown", "id": "4db4f47b", "metadata": {}, "source": [ "And we can verify, using the full text of Federalist No. 83, that the two-word\n", "phrase *by jury* appears 36 times in the text. Setting these appearances aside, we can\n", "further check that the rate of *by* (without *by jury*) is about 7 per 1,000 words." ] }, { "cell_type": "code", "execution_count": 15, "id": "0db8d3eb", "metadata": { "tags": [ "remove-cell" ] }, "outputs": [], "source": [ "# HIDE THIS CELL (data check)\n", "assert df_known.loc[df_known['by'] > 30, 'by'].index == 83" ] }, { "cell_type": "code", "execution_count": 16, "id": "816fc591", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "In Federalist No. 83 (by Hamilton), without \"by jury\", \"by\" occurs 7 times per 1,000 words on average.\n" ] } ], "source": [ "with open('data/federalist-83.txt') as infile:\n", " text = infile.read()\n", "\n", "# a regular expression here would be more robust\n", "by_jury_count = text.count(' by jury')\n", "by_count = text.count(' by ')\n", "word_count = len(text.split()) # crude word count\n", "by_rate = 1000 * (by_count - by_jury_count) / word_count\n", "\n", "print('In Federalist No. 83 (by Hamilton), without \"by jury\", '\n", " f'\"by\" occurs {by_rate:.0f} times per 1,000 words on average.')" ] }, { "cell_type": "markdown", "id": "c5797aee", "metadata": {}, "source": [ "In the block above, we read in a file containing the text of Federalist No. 83 and\n", "count the number of times *by jury* occurs and the number of times *by*\n", "(without *jury*) occurs. We then normalize the count of *by* (without *jury*)\n", "to the rate per 1,000 words for comparison with the rate observed elsewhere.\n", "\n", "While it would be convenient to adopt the approach we took with *upon* and apply\n", "it to the present case, unfortunately we cannot. The word *by* appears in every\n", "one of the essays with an established author; the approach we used earlier will\n", "not work. That is, if we expect both authors to use *by* at least once per 1,000\n", "words, the probability of the author being Hamilton given that *by* occurs at\n", "least once is very high and the probability of the author being Madison given\n", "that *by* occurs at least once is also very high. In such a situation---when\n", "$\\Pr(E|H) \\approx \\Pr(E|\\neg H)$---Bayes's rule guarantees we will learn little:\n", "$\\Pr(H|E) \\approx P(H)$. In order to make use of information about the frequency\n", "of *by* in the unknown essay, we need a more nuanced model of each author's\n", "tendency to use *by*. We need to have a precise answer to questions such as the\n", "following: if we pick 1,000 words at random from Madison's writing, how likely\n", "are we to find 6 instances of *by* among them? More than 12 instances? How\n", "surprising would it be to find 40 or more?\n", "\n", "#### Negative binomial distribution\n", "\n", "We need a model expressive enough to answer all these questions. This model must be\n", "capable of saying that among 1,000 words chosen at random from the known essays by an\n", "author, the probability of finding $x$ instances of *by* is a specific probability. In\n", "symbols we would use the symbol $X$ to denote the \"experiment\" of drawing 1,000 words at\n", "random from an essay and reporting the number of *by* instances and write lowercase $x$\n", "for a specific realization of the experiment. This lets us write $p = \\Pr(X = x|H) = f(x)$\n", "to indicate the probability of observing a rate of $x$ *by* tokens per 1,000 words given\n", "that the text is written by Hamilton. Whatever function, $f$, we choose to model $X$\n", "(called a *random variable*), we will need it to respect the axioms of probability. For\n", "example, following axiom 1, it must not report a probability less than zero or greater\n", "than 1. There are many functions (called \"probability mass\n", "functions\") which will do the job. Some will be more faithful than others in\n", "capturing the variability of rates of *by* which we observe (or feel confident we will\n", "observe). That is, some will be better aligned with our degrees of belief in observing the\n", "relevant rate of *by* given that Hamilton is the author. Candidate distributions include\n", "the Poisson distribution, the binomial\n", "distribution, and the negative binomial distribution. We\n", "will follow Mosteller and Wallace and use the negative binomial distribution.\n", "```{margin}\n", "Mosteller and Wallace\n", "justify their choice of the negative binomial distribution at length\n", "({cite:t}`mosteller1964inference`, 28--35).\n", "```\n", "\n", "$X$, a *random variable*, is said to be distributed according\n", "to a negative binomial distribution with parameters $\\alpha$ and $\\beta$ if\n", "\n", "$$\\Pr(X = x \\,|\\, \\alpha, \\beta) =\n", "\\binom{x + \\alpha - 1}{\\alpha - 1}\n", "\\,\n", "\\left( \\frac{\\beta}{\\beta+1} \\right)^{\\!\\alpha}\n", "\\,\n", "\\left( \\frac{1}{\\beta + 1} \\right)^{\\!x} \\!.\n", "$$\n", "\n", "```{admonition} Random variable\n", "A *random variable* is defined as a function from a *sample space* into the real numbers. A random variable, $X$, intended to represent whether the toss of a coin landed \"heads\" would be associated with the sample space $S = \\{H, T\\}$, where $H$ and $T$ indicate the two possible outcomes of the toss ($H$ for \"heads\" and $T$ for \"tails\"). Each of these potential outcomes is associated with a probability. In this case, $\\Pr(H)$ and $\\Pr(T)$ might be equal to 0.5. The random variable $X$ takes on numeric values; here it would take on the value 1 if the outcome of the random experiment (i.e., \"flipping a coin\") was $H$, otherwise the random variable would take on the value 0. That is, we observe $X = 1$ if and only if the outcome of the random experiment is $H$. The probability function $\\Pr$ associated with the sample space induces a probability function associated with $X$ which we could write $\\Pr_X$. This function must satisfy the axioms of probability. In symbols, it would be written $\\Pr_X(X = x_i) = \\Pr(\\{s_j \\in S : X(s_j) = x_i\\})$ ({cite:t}`casella2001statistical`, section 1.1-1.4).\n", "```\n", "\n", "```{margin}\n", "When dealing with continuous random variables,\n", "we need \"probability density functions\" or \"pdfs\". Working with these, however, requires\n", "additional care as the probability of a point is always\n", "zero---just as the mass of a physical object at a point is also zero.\n", "```\n", "This function $\\Pr(X = x \\,|\\, \\alpha, \\beta)$ is a probability mass\n", "function (abbreviated \"pmf\"). If a\n", "random variable follows this distribution we may also write $X \\sim\n", "\\distro{NegBinom}(\\alpha,\\beta)$ and say that $X$ is distributed according\n", "to a negative binomial distribution. \n", "```{margin}\n", "The\n", "parameterization of the negative binomial distribution follows the\n", "parameterization described in {cite:t}`gelman2003bayesian`. Mosteller and Wallace use a\n", "different parameterization of the negative binomial distribution. The choice of\n", "parameterization does not affect results.\n", "```\n", "Given a sequence of draws from such a\n", "random variable, we would find that the sequence has mean\n", "$\\frac{\\alpha}{\\beta}$ and variance $\\frac{\\alpha}{\\beta^2} (\\beta + 1)$. The derivation of these results from\n", "the negative binomial pmf above is a subject covered in most probability courses\n", "and introductory textbooks.\n", "\n", "For example, if $X \\sim \\distro{NegBinom}(5, 1)$ then the probability of\n", "observing $X = 6$ is, according to the pmf above, approximately 10 percent. The\n", "probability of observing $X = 14$ is considerably lower, 0.5 percent. We can\n", "verify these calculations by translating the negative binomial pmf stated above\n", "into Python code. The following function, ``negbinom_pmf()``, makes use of one\n", "possibly unfamiliar function, ``scipy.special.binom(a, b)`` which calculates,\n", "as the name suggests, the binomial coefficient $\\binom{a}{b}$." ] }, { "cell_type": "code", "execution_count": 17, "id": "df11771e", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Pr(X = 6): 0.1025390625\n", "Pr(X = 14): 0.00583648681640625\n" ] } ], "source": [ "import scipy.special\n", "\n", "def negbinom_pmf(x, alpha, beta):\n", " \"\"\"Negative binomial probability mass function.\"\"\"\n", " # In practice this calculation should be performed on the log\n", " # scale to reduce the risk of numeric underflow.\n", " return (\n", " scipy.special.binom(x + alpha - 1, alpha - 1)\n", " * (beta / (beta + 1)) ** alpha\n", " * (1 / (beta + 1)) ** x\n", " )\n", "\n", "print('Pr(X = 6):', negbinom_pmf(6, alpha=5, beta=1))\n", "print('Pr(X = 14):', negbinom_pmf(14, alpha=5, beta=1))" ] }, { "cell_type": "markdown", "id": "bc3c65db", "metadata": {}, "source": [ "Further discussion of discrete distributions and probability mass functions is\n", "not necessary to appreciate the immediate usefulness of the negative binomial\n", "distribution as a tool which provides answers to the questions asked above.\n", "With suitably chosen parameters $\\alpha$ and $\\beta$, the negative binomial\n", "distribution (and its associated probability mass function) provides precise\n", "answers to questions having the form we are interested in, questions such as\n", "the following: how likely is it for a 1,000-word essay by Hamilton to\n", "contain 6 instances of *by*? Moreover, the negative binomial distribution can\n", "provide plausible answers to such questions, where plausibility is defined with\n", "reference to our degrees of belief.\n", "\n", "How well the answers align with our beliefs about plausible rates of *by*\n", "depends on the choice of the parameters $\\alpha$ and $\\beta$. It is not\n", "difficult to find appropriate parameters, as the graphs below will make clear.\n", "The graphs show the observed rates of *by* from Hamilton's essays alongside\n", "simulated rates drawn from negative binomial distributions with three different\n", "parameter settings. How we arrived at these precise parameter settings is\n", "addressed in the Appendix {ref}`appendix-intro-probability-fitting-nbd`.\n", "\n", "The first plot bewlow will show the empirical distribution of the *by* rates we\n", "observe in Hamilton's (non-disputed) essays. These observed rates are all the\n", "information we have about Hamilton's tendency to use *by*, so we will choose a\n", "negative binomial distribution which roughly expresses the same information." ] }, { "cell_type": "code", "execution_count": 18, "id": "7ffc850a", "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "filenames": { "image/png": "/Users/folgert/projects/hda/_build/jupyter_execute/intro-probability/notebook_33_0.png" } }, "output_type": "display_data" } ], "source": [ "df_known[df_known['AUTHOR'] == 'HAMILTON']['by'].plot.hist(\n", " range=(0, 35), density=True, rwidth=0.9);" ] }, { "cell_type": "markdown", "id": "371cdd69", "metadata": {}, "source": [ "" ] }, { "cell_type": "code", "execution_count": 19, "id": "9f4984c3", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "count 51.000000\n", "mean 7.019608\n", "std 4.925404\n", "min 1.000000\n", "25% 5.000000\n", "50% 6.000000\n", "75% 9.000000\n", "max 34.000000\n", "Name: by, dtype: float64" ] }, "execution_count": 19, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df_known[df_known['AUTHOR'] == 'HAMILTON']['by'].describe()" ] }, { "cell_type": "markdown", "id": "6c622ffc", "metadata": {}, "source": [ "Now that we know what we're looking for -- a distribution with plenty of mass\n", "between 1 and 10 and minimal support for observations greater than 40 -- we can\n", "look at various negative binomial distributions and see if any will work as\n", "approximations of our degrees of belief about Hamilton's tendency to use *by*\n", "around 1787." ] }, { "cell_type": "code", "execution_count": 20, "id": "70f18239", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans.\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "findfont: Generic family 'sans-serif' not found because none of the following families were found: \"Roboto Condensed Regular\"\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "filenames": { "image/png": "/Users/folgert/projects/hda/_build/jupyter_execute/intro-probability/notebook_37_2.png" } }, "output_type": "display_data" } ], "source": [ "import itertools\n", "import matplotlib.pyplot as plt\n", "\n", "x = np.arange(60)\n", "alphas, betas = [5, 6.5, 12, 16], [1.5, 0.5, 0.7]\n", "params = list(itertools.product(alphas, betas))\n", "pmfs = [negbinom_pmf(x, alpha, beta) for alpha, beta in params]\n", "\n", "fig, axes = plt.subplots(4, 3, sharey=True, figsize=(10, 8))\n", "axes = axes.flatten()\n", "\n", "for ax, pmf, (alpha, beta) in zip(axes, pmfs, params):\n", " ax.bar(x, pmf)\n", " ax.set_title(fr'$\\alpha$ = {alpha}, $\\beta$ = {beta}')\n", "plt.tight_layout();" ] }, { "cell_type": "markdown", "id": "34a0b6f7", "metadata": {}, "source": [ "\n", "\n", "The parameterization of $\\alpha = 5$ and $\\beta = 0.7$ is one of the models that\n", "(visually) resembles the empirical distribution. Indeed, if we accept\n", "$\\distro{NegBinom}(5, 0.7)$ as modeling our degrees of belief about Hamilton's\n", "use of *by*, the distribution has the attractive feature of predicting the same\n", "average rate of *by* as we observe in the Hamilton essays in the corpus. We can\n", "verify this by simulating draws from the distribution and comparing the summary\n", "statistics of our sampled values with the observations we have encountered so\n", "far." ] }, { "cell_type": "code", "execution_count": 21, "id": "418be813", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "count 10000.00000\n", "mean 7.07620\n", "std 4.11972\n", "min 0.00000\n", "25% 4.00000\n", "50% 6.00000\n", "75% 9.00000\n", "max 29.00000\n", "dtype: float64" ] }, "execution_count": 21, "metadata": {}, "output_type": "execute_result" } ], "source": [ "def negbinom(alpha, beta, size=None):\n", " \"\"\"Sample from a negative binomial distribution.\n", "\n", " Uses `np.random.negative_binomial`, which makes use of a\n", " different parameterization than the one used in the text.\n", " \"\"\"\n", " n = alpha\n", " p = beta / (beta + 1)\n", " return np.random.negative_binomial(n, p, size)\n", "\n", "samples = negbinom(5, 0.7, 10000)\n", "# put samples in a pandas Series in order to calculate summary statistics\n", "pd.Series(samples).describe()" ] }, { "cell_type": "markdown", "id": "c4de07f4", "metadata": {}, "source": [ "The negative binomial distribution with fixed\n", "parameters---$\\alpha = 5$ and $\\beta = 0.7$ were used in the code block above---is an\n", "oracle which answers the questions we have. For any observed value of the rate of *by* in\n", "a disputed essay we can provide an answer to the question, \"Given that Hamilton is the\n", "author of the disputed essay, what is the probability of observing a rate $x$ of *by*?\"\n", "The answer is $\\Pr(X = x|\\alpha, \\beta)$ for fixed values of $\\alpha$ and $\\beta$,\n", "something we can calculate directly with the function ``negbinom_pmf(x, alpha, beta)``.\n", "Now that we have a way of calculating the probability of *by* under the hypothesis that\n", "Hamilton is the author, we can calculate a quantity that plays the same role that\n", "$\\Pr(E|H)$ played in the previous section. This will allow us to use Bayes's rule to calculate our degree of belief in the claim that\n", "Hamilton wrote a disputed essay, given an observed rate of *by* in the essay.\n", "\n", "Now let us consider some plausible models of Madison's use of the word *by*. In\n", "order to use Bayes's rule we will need to be able to characterize how plausible\n", "different rates of *by* are, given that Madison is the author." ] }, { "cell_type": "code", "execution_count": 22, "id": "6b7d3072", "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "filenames": { "image/png": "/Users/folgert/projects/hda/_build/jupyter_execute/intro-probability/notebook_41_0.png" } }, "output_type": "display_data" } ], "source": [ "df_known[df_known['AUTHOR'] == 'MADISON']['by'].plot.hist(\n", " density=True, rwidth=0.9, range=(0, 35) # same scale as with Hamilton\n", ");" ] }, { "cell_type": "markdown", "id": "818eb1ef", "metadata": {}, "source": [ "" ] }, { "cell_type": "code", "execution_count": 23, "id": "f4341b29", "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "filenames": { "image/png": "/Users/folgert/projects/hda/_build/jupyter_execute/intro-probability/notebook_43_0.png" } }, "output_type": "display_data" } ], "source": [ "x = np.arange(60)\n", "alphas, betas = [100, 50, 28, 10], [23, 4, 1.3]\n", "params = list(itertools.product(alphas, betas))\n", "pmfs = [negbinom_pmf(x, alpha, beta) for alpha, beta in params]\n", "\n", "fig, axes = plt.subplots(4, 3, sharey=True, figsize=(10, 8))\n", "axes = axes.flatten()\n", "\n", "for ax, pmf, (alpha, beta) in zip(axes, pmfs, params):\n", " ax.bar(x, pmf)\n", " ax.set_title(fr'$\\alpha$ = {alpha}, $\\beta$ = {beta}')\n", "plt.tight_layout()" ] }, { "cell_type": "markdown", "id": "0208b675", "metadata": {}, "source": [ "\n", "\n", "If $\\alpha$ is 50 and $\\beta$ is 4, the model resembles the empirical\n", "distribution of the rates of *by* in the Madison essays. Most values are between\n", "5 and 25 and the theoretical mean of the negative binomial distribution is\n", "similar to the empirical mean. Compare the two (idealized) models of the use of\n", "*by* with the rates of *by* we have observed in the essays with known\n", "authorship.\n", "\n", "" ] }, { "cell_type": "code", "execution_count": 24, "id": "c58066e6", "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "filenames": { "image/png": "/Users/folgert/projects/hda/_build/jupyter_execute/intro-probability/notebook_45_0.png" } }, "output_type": "display_data" } ], "source": [ "authors = ('HAMILTON', 'MADISON')\n", "alpha_hamilton, beta_hamilton = 5, 0.7\n", "alpha_madison, beta_madison = 50, 4\n", "\n", "# observed\n", "fig, axes = plt.subplots(2, 1, sharex=True, sharey=True)\n", "df_known.groupby('AUTHOR')['by'].plot.hist(\n", " ax=axes[0], density=True, range=(0, 35), rwidth=0.9, alpha=0.6,\n", " title='Hamilton v. Madison (observed)', legend=True)\n", "\n", "# model\n", "simulations = 10000\n", "for author, (alpha, beta) in zip(authors, [(alpha_hamilton, beta_hamilton),\n", " (alpha_madison, beta_madison)]):\n", " pd.Series(negbinom(alpha, beta, size=simulations)).plot.hist(\n", " label=author, density=True, rwidth=0.9, alpha=0.6, range=(0, 35), ax=axes[1])\n", "axes[1].set_xlim((0, 40))\n", "axes[1].set_title('Hamilton v. Madison (model)')\n", "axes[1].legend()\n", "plt.tight_layout();" ] }, { "cell_type": "markdown", "id": "99ffd7ed", "metadata": {}, "source": [ "\n", "\n", "With the negative binomial distribution available as a rough model for our\n", "beliefs about each author's use of *by*, we have all that is required to use\n", "Bayes's rule and assess how observing a given rate of *by* in a disputed essay\n", "should persuade us of the plausibility of Hamilton or Madison being the author.\n", "\n", "Imagine we are presented with one of the disputed essays and observe that\n", "the word *by* occurs 14 times per 1,000 words---the rate observed, in fact, in\n", "No. 57. A casual inspection of the rates for essays with known authorship tells\n", "us that 14 instances of *by* would be rather high for Hamilton (9 per 1,000 on\n", "average in our samples); 14 is closer to the expected rate for Madison. So we\n", "should anticipate that the calculation suggested by Bayes' rule would increase\n", "the odds favoring Madison. Using $\\Pr(H) = \\Pr(\\neg H) = 0.5$ as before, we use\n", "the probability mass function for the negative binomial distribution to\n", "calculate the probability of observing a rate of 14 per 1,000 in an essay\n", "authored by Hamilton and the probability of observing a rate of 14 per 1,000 in\n", "an essay authored by Madison (probabilities analogous to $\\Pr(E|H)$ and $\\Pr(E|\\neg H)$ in the\n", "previous section). Instead of $E$ we will write $X$ for the observed rate of\n", "*by*. So $\\Pr(X = 14|H) = \\Pr(X = x|\\alpha_{Hamilton},\\beta_{Hamilton}) = 0.022$ and\n", "$\\Pr(X = 14|\\neg H) = \\Pr(X = x|\\alpha_{Madison}, \\beta_{Madison}) = 0.087$. The way\n", "we find these values is to use the function ``negbinom_pmf()``, which we defined\n", "above." ] }, { "cell_type": "code", "execution_count": 25, "id": "d7610452", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "0.021512065936254765\n" ] } ], "source": [ "likelihood_hamilton = negbinom_pmf(14, alpha_hamilton, beta_hamilton)\n", "print(likelihood_hamilton)" ] }, { "cell_type": "code", "execution_count": 26, "id": "c1552b86", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "0.08742647980678281\n" ] } ], "source": [ "likelihood_madison = negbinom_pmf(14, alpha_madison, beta_madison)\n", "print(likelihood_madison)" ] }, { "cell_type": "markdown", "id": "07d9f343", "metadata": {}, "source": [ "$$\\begin{aligned}\n", "\\Pr(H|X = 14, \\vec\\theta) &= \\frac{\\Pr(X = 14|H)\\Pr(H)}{\\Pr(X = 14|H)\\Pr(H) + \\Pr(X = 14|\\neg H)\\Pr(\\neg H)} \\\\\n", " &= \\frac{\\Pr(X = 14|\\alpha_{Hamilton}, \\beta_{Hamilton})\\Pr(H)}{\\Pr(X = 14|\\alpha_{Hamilton}, \\beta_{Hamilton})\\Pr(H) + \\Pr(X = 14|\\alpha_{Madison}, \\beta_{Madison})\\Pr(\\neg H)} \\\\\n", " &\\approx \\frac{(0.02)(0.5)}{(0.02)(0.5) + (0.08)(0.5)} \\\\\n", " &\\approx 0.2 \\\\\n", "\\end{aligned}$$" ] }, { "cell_type": "code", "execution_count": 27, "id": "996100ce", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "0.19746973662561154\n" ] } ], "source": [ "pr_hamilton = likelihood_hamilton * 0.5 / (likelihood_hamilton * 0.5 + likelihood_madison * 0.5)\n", "print(pr_hamilton)" ] }, { "cell_type": "markdown", "id": "e7e459ef", "metadata": {}, "source": [ "In the code above we've used the term *likelihood* to refer to the probability\n", "of the evidence given the hypothesis, $\\Pr(E|H)$.\n", "\n", "Bayes's rule tells us how to update our beliefs about whether or not Hamilton\n", "wrote this hypothesized disputed essay about which we have only learned the rate\n", "of *by* instances is 14 per 1,000. In such a case the odds turn against Hamilton\n", "being the author, approximately 4 to 1 in favor of Madison. The rate of *by* in\n", "the unknown essay is just one piece of evidence which we might consider in\n", "assessing how much plausibility we assign to the claim \"Hamilton wrote paper\n", "No. 52\". Mosteller and Wallace consider the rates of 30 words (*by* is in the\n", "group of \"final words\") in their analysis ({cite:t}`mosteller1964inference`, 67--68).\n", "\n", "(sec-intro-probability-further-reading)=\n", "## Further Reading\n", "\n", "This chapter introduced Bayesian inference, one technique for learning from\n", "experience. Among many possible approaches to learning from observation,\n", "Bayesian inference provides a specific recipe for using probabilities to\n", "describe degrees of belief and for updating degrees of belief based on\n", "observation. Another attractive feature of Bayesian inference is its\n", "generality. Provided we can come up with a description of our prior degree of\n", "belief in an event occurring, as well as a description of how probable some observation would be under\n", "various hypotheses about the event,\n", "Bayes's rule provides us with a recipe for updating our degree of belief in the\n", "event being realized after taking into consideration the observation. To\n", "recall the example that we concluded with, Bayesian inference provides us with\n", "a principled way of arriving at the claim that \"it is very likely that Madison\n", "(rather than Hamilton) wrote Federalist No. 62\" given observed rates of word\n", "usage in Federalist No. 62. This is a claim that historians writing before\n", "1950 had no way of substantiating. Thanks to Bayesian inference and the work of\n", "Mosteller and Wallace, the evidence and procedure supporting this claim are\n", "accessible to everyone interested in this case of disputed authorship.\n", "\n", "For essential background reading related to this chapter, we recommend {cite:t}`grinstead2012introduction`\n", "which provides an introduction to discrete and continuous probability. Their book is\n", "published by the American Mathematical Society and is available [online](https://math.dartmouth.edu/~prob/prob/prob.pdf) at. For those interested in further reading\n", "related to the topics addressed in this chapter, we recommend an introductory text on Bayesian inference. Those with fluency in single-variable\n", "calculus and probability will be well served by {cite:t}`hoff2009first`. While {cite:t}`hoff2009first` uses the R programming language for\n", "performing computation and visualizing results, the code provided may be\n", "translated into Python without much effort.\n", "\n", "## Excercises\n", "### Easy\n", "1. Which of the following terms is used to denote a prior belief? a) $\\Pr(E|H)$, b)\n", " $\\Pr(H|E)$, or c) $\\Pr(H)$.\n", "2. Which of the following terms is used to describe the likelihood of an observation given a hypothesis? a) $\\Pr(E|H)$, b)\n", " $\\Pr(E)$, or c) $\\Pr(H|E)$.\n", "3. Recall the example about Pynchon from the introduction to this chapter. Suppose we improve our\n", " stylometric test to accurately identify a novel as being written by Pynchon from 90\n", " percent to 99 percent of the time. The false positive rate also decreased and now\n", " equals 0.1 percent. The probability that a novel was written by Pynchon is still 0.001\n", " percent. Suppose another text tests positive on our stylometric test. What is the\n", " probability that the text was written by Pynchon?\n", "\n", " ```{margin}\n", " Each essay is not a\n", " sequence of 1,000 words. An essay is a sequence of words of fixed length\n", " (assuming we can agree on word-splitting rules). The writing samples from\n", " Madison and Hamilton tend to be about 2,000 words, on average, and the words\n", " of interest are common so the consequences of this infidelity to what we\n", " know to be the case will be limited.\n", " ```\n", "4. Mosteller and Wallace describe Madison's and Hamilton's word usage in terms of\n", " frequency per 1,000 words. While most essays were longer---typically\n", " between 2,000 and 3,000 words---pretending as if each document were 1,000\n", " words and contained a fixed number of occurrences of the words of interest\n", " allows us to compare texts of different lengths. Inaccuracies introduced by\n", " rounding will not be consequential in this case. Calculate the frequency per 1,000\n", " words of *upon*, *the*, and *enough*.\n", "\n", "### Moderate\n", "Mosteller and Wallace started their investigation of the authorship of the\n", "disputed essays in *The Federalist Papers* by focusing on a handful of words, which, closely\n", "reading the essays, had revealed as distinctive: *while*, *whilst*, *upon*,\n", "*enough*. Focus on the word *enough*. Suppose you are about to inspect one\n", "of the disputed essays and see if *enough* appears.\n", "\n", "1. How many times does the word *enough* occur at least once in essays by Madison? How\n", " many times does the word occur at least once in essays by Hamilton?\n", "2. Establish values for $\\Pr(H)$, $\\Pr(E|H)$, and $\\Pr(E|\\neg H)$ that you find\n", " credible. ($\\Pr(E|H)$ here is the probability that the word *enough* appears in a\n", " disputed essay when Hamilton, rather than Madison, is the author.)\n", "3. Suppose you learn that *enough* appears in the disputed essay. How does your belief\n", " about the author change?\n", "\n", "### Challenging\n", "1. Consider the rate at which the word *of* occurs in texts with known authorship. If you\n", " were to use a binomial distribution (not a negative binomial distribution) to model\n", " each author's use of the word *of* (expressed in frequency per 1,000 words), what\n", " value would you give to the parameter $\\theta$ associated with Hamilton? And with\n", " Madison?\n", "2. Working with the parameter values chosen above, suppose you observe a disputed essay\n", " with a rate of 8 *of*s per 1,000 words. Does this count as evidence in favor of\n", " Madison being the author or as evidence in favor of Hamilton being the author?\n", "\n", "(appendix-intro-probability)=\n", "## Appendix\n", "\n", "(appendix-intro-probability-bayes-rule)=\n", "### Bayes's rule\n", "\n", "In order to derive Bayes's rule we first begin with the third axiom of\n", "probability:\n", "\n", "$$\\Pr(A \\and B|C) = \\Pr(B|C) \\Pr(A|B \\and C)$$\n", "\n", "If we let $C$ be an event which encompasses all possible events ($A \\and C = A$)\n", "we have a simpler statement about a conditional probability:\n", "\n", "$$\\Pr(A|B) = \\frac{\\Pr(A \\and B)}{\\Pr(B)}$$\n", "\n", "Replacing $A$ and $B$ with $H_j$ and $E$, respectively, we arrive at an initial\n", "form of Bayes's rule:\n", "\n", "$$\\begin{aligned}\n", "\\Pr(H_j|E) &= \\frac{\\Pr(E \\and H_j)}{\\Pr(E)} \\\\\n", " &= \\frac{\\Pr(H_j|E)\\Pr(E)}{\\Pr(E)} \\\\\n", "\\end{aligned}$$\n", "\n", "```{margin}\n", "The rule of marginal probability for discrete random\n", "variables is the following:\n", "\n", "$$\\begin{aligned}\n", "\\Pr(X = x) &= \\sum_{y \\in \\mathcal{Y}} \\Pr(X = x, Y = y) \\\\\n", " &= \\sum_{y \\in \\mathcal{Y}} \\Pr(X = x| Y = y)\\Pr(Y = y)\n", "\\end{aligned}$$\n", "\n", "where $\\mathcal{Y}$ is the set of values which $Y$ may take. In prose, the rule for\n", " marginal probability tells us how to calculate the marginal distribution,\n", " $\\Pr(X = x)$, if we know the joint distribution $\\Pr(X = x, Y = y)$.\n", " (Appealing to the third axiom of probability, we can observe that the joint\n", " distribution of $X$ and $Y$ is equal to the product of the conditional\n", " distribution of X given Y times the marginal distribution of Y, $\\Pr(X = x |\n", " Y = y)\\Pr(Y = y)$.)\n", "```\n", "The denominator, $\\Pr(E)$ may be unpacked by using the rule of marginal\n", "probability:\n", "\n", "$$\\begin{aligned}\n", "\\Pr(E) &= \\sum_{k=1}^K \\Pr(E \\and H_k) \\\\\n", " &= \\sum_{k=1}^K \\Pr(E|H_k)\\Pr(H_k) \\\\\n", "\\end{aligned}$$\n", "\n", "Replacing $\\Pr(E)$ in the initial statement we have the final, familiar form of\n", "Bayes's rule:\n", "\n", "$$\\Pr(H_j|E) = \\frac{\\Pr(E|H_j)\\Pr(H_j)}{\\sum_{k=1}^K \\Pr(E|H_k)\\Pr(H_k)}$$\n", "\n", "(appendix-intro-probability-fitting-nbd)=\n", "### Fitting a negative binomial distribution\n", "\n", "Finding values for $\\alpha$ and $\\beta$ which maximize the sampling probability\n", "(as a function of $\\alpha$ and $\\beta$) is tricky. Given knowledge of the sample\n", "mean $\\bar x$ and the value of $\\alpha$, the value of $\\beta$ which maximizes the\n", "sampling probability (as a function of $\\beta$) can be found by finding the\n", "critical points of the function. (The answer is $\\hat \\beta = \\frac{\\alpha}{\\bar\n", "x}$ and the derivation is left as an exercise for the reader.) Finding a good\n", "value for $\\alpha$ is challenging. The path of least resistance here is\n", "numerical optimization. Fortunately, the package SciPy comes with a general\n", "purpose function for optimization which tends to be reliable and requires\n", "minimal configuration:" ] }, { "cell_type": "code", "execution_count": 28, "id": "20497410", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "count 17.000000\n", "mean 11.352941\n", "std 10.024602\n", "min 4.000000\n", "25% 6.000000\n", "50% 8.000000\n", "75% 13.000000\n", "max 47.000000\n", "dtype: float64" ] }, "execution_count": 28, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# `x` is a sample of Hamilton's rates of 'by' (per 1,000 words)\n", "x = np.array([13, 6, 4, 8, 16, 9, 10, 7, 18, 10, 7, 5, 8, 5, 6, 14, 47])\n", "pd.Series(x).describe()" ] }, { "cell_type": "code", "execution_count": 29, "id": "2308fbe3", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "2.9622164213528452 0.260920617424862\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "/var/folders/s8/1h01d3xx303c51_q78_b9gjw0000gn/T/ipykernel_48099/183489177.py:14: RuntimeWarning: divide by zero encountered in log\n", " return -1 * np.sum(np.log(negbinom_pmf(x, alpha, beta)))\n" ] } ], "source": [ "import scipy.optimize\n", "import scipy.special\n", "\n", "# The function `negbinom_pmf` is defined in the text.\n", "\n", "\n", "def estimate_negbinom(x):\n", " \"\"\"Estimate the parameters of a negative binomial distribution.\n", "\n", " Maximum-likelihood estimates of the parameters are calculated.\n", " \"\"\"\n", " def objective(x, alpha):\n", " beta = alpha / np.mean(x) # MLE for beta has closed-form solution\n", " return -1 * np.sum(np.log(negbinom_pmf(x, alpha, beta)))\n", " alpha = scipy.optimize.minimize(\n", " lambda alpha: objective(x, alpha), x0=np.mean(x), bounds=[(0, None)], tol=1e-7).x[0]\n", " return alpha, alpha / np.mean(x)\n", "\n", "\n", "alpha, beta = estimate_negbinom(x)\n", "print(alpha, beta)" ] } ], "metadata": { "jupytext": { "cell_metadata_filter": "-all", "formats": "md:myst", "text_representation": { "extension": ".md", "format_name": "myst", "format_version": 0.13, "jupytext_version": "1.10.3" } }, "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.9.13" }, "source_map": [ 14, 32, 36, 114, 120, 449, 457, 467, 476, 485, 490, 503, 506, 517, 521, 525, 533, 540, 562, 564, 568, 570, 664, 675, 682, 685, 705, 707, 713, 719, 731, 826, 841, 867, 870, 874, 876, 884, 900, 913, 927, 946, 950, 954, 967, 980, 1001, 1027, 1032, 1035, 1044, 1047, 1212, 1218 ] }, "nbformat": 4, "nbformat_minor": 5 }