DOCSLIB.ORG

Lecture 2: Conjugate Priors

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text

Load more

CS598JHM: Advanced NLP (Spring ʼ10) The binomial distribution If p is the probability of heads, the probability of getting exactly k heads in n independent yes/no trials is given by Lecture 2: the binomial distribution Bin(n,p): Conjugate priors n k n k P (k heads) = p (1 p) − k − n! k n k = p (1 p) − k!(n k)! − − Julia Hockenmaier Expectation E(Bin(n,p)) = np [email protected] 3324 Siebel Center Variance var(Bin(n,p)) = np(1-p) http://www.cs.uiuc.edu/class/sp10/cs598jhm CS598JHM: Advanced NLP 2 Binomial likelihood Parameter estimation What distribution does p (probability of heads) have, Given a set of data D=HTTHTT, what is the probability ! of given that the data D consists of #H heads and #T tails? heads? Likelihood L(e;D=(#Heads,#Tails)) for binomial distribution 0.007 - Maximum likelihood estimation (MLE): L(e,(5,5)) L(e,(3,7)) Use the ! which has the highest likelihood P(D| !). 0.006 L(e,(2,8)) P (x = H D)=P (x = H θ)withθ = arg max P (D θ) 0.005 | | θ | 0.004 - Bayesian estimation: 0.003 Compute the expectation of ! given D: 1 0.002 P (x = H D)= P (x = H θ)P (θ D)dθ = E[θ D] | 0 | | | 0.001 0 0 0.2 0.4 0.6 0.8 1 CS598JHM: Advancede NLP CS598JHM: Advanced NLP 3 4 Maximum likelihood estimation Bayesian statistics - Data D provides evidence for or against our beliefs. - Maximum likelihood estimation (MLE): We update our belief ! based on the evidence we see: find ! which maximizes likelihood P(D | !). Prior Likelihood P (θ)P (D θ) P (θ D)= | | P (θ)P (D θ)dθ θ∗ = arg max P (D θ) θ | Posterior | = arg max θH (1 θ)T Marginal Likelihood (=P(D)) θ − H = H + T CS598JHM: Advanced NLP CS598JHM: Advanced NLP 5 6 Bayesian estimation In search of a prior... Given a prior P(!) and a likelihood P(D|!), We would like something of the form: what is the posterior P(! |D)? P (θ) θa(1 θ)b How do we choose the prior P(!)? ∝ − - The posterior is proportional to prior x likelihood: But -- this looks just like the binomial: P(! |D)! P(!) P(D|!) n k n k P (k heads) = p (1 p) − k − - The likelihood of a binomial is: H T P(D|!) = ! (1-!) n! k n k = p (1 p) − k!(n k)! − - If prior P(!) is proportional to powers of ! and (1-!), − posterior will also be proportional to powers of ! and (1-!): P(!) ! ! a(1-!)b …. except that k is an integer and ! is a real with 0<! <1. " P(! |D)! ! a(1-!)b !H(1-!)T = !a+H(1-!)b+T CS598JHM: Advanced NLP CS598JHM: Advanced NLP 7 8 The Gamma function The Gamma function The Gamma function "(x) is the generalization of the K(x) function factorial x! (or rather (x-1)!) to the reals: 25 ∞ α 1 x 20 Γ(α)= x − e− dx for α > 0 0 15 For x >1, "(x) = (x-1)"(x-1). 10 For positive integers, "(x) = (x-1)! 5 0 0 1 2 3 4 5 CS598JHM: Advanced NLP CS598JHM: Advanced NLP 9 10 The Beta distribution Beta(",#) with " >1, # >1 A random variable X (0 < x < 1) has a Beta distribution with Unimodal (hyper)parameters " (" > 0) and # (# > 0) if X has a continuous Beta Distribution Beta(_, `) distribution with probability density function 7 Beta(1.5,1.5) Beta(3,1.5) 6 Beta(3,3) Γ(α + β) α 1 β 1 Beta(20,20) P (x α, β)= x − (1 x) − Beta(3,20) | Γ(α)Γ(β) − 5 4 The first term is a normalization factor (to obtain a distribution) 3 1 α 1 β 1 Γ(α + β) 2 x − (1 x) − dx = 0 − Γ(α)Γ(β) 1 Expectation: α α+β 0 0 0.2 0.4 0.6 0.8 1 ! CS598JHM: Advanced NLP CS598JHM: Advanced NLP 11 12 Beta(",#) with " <1, # <1 Beta(",#) with "=# U-shaped Symmetric. #=$=1: uniform Beta Distribution Beta(_, `) Beta Distribution Beta(_, `) 6 3.5 Beta(0.1,0.1) Beta(0.1,0.1) Beta(0.1,0.5) Beta(1,1) 5 Beta(0.5,0.5) 3 Beta(2,2) 2.5 4 2 3 1.5 2 1 1 0.5 0 0 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 ! ! CS598JHM: Advanced NLP CS598JHM: Advanced NLP 13 14 Beta(",#) with "<1, # >1 Beta(",#) with " = 1, # >1 Strictly decreasing # = 1, 1< $ < 2: strictly concave. Beta Distribution Beta(_, `) # = 1, $ = 2: straight line 8 # = 1, $ > 2: strictly convex Beta(0.1,1.5) Beta(0.5,1.5) Beta Distribution Beta(_, `) 7 Beta(0.5,2) 3.5 Beta(1,1.5) 6 Beta(1,2) 3 Beta(1,3) 5 2.5 4 2 3 1.5 2 1 1 0 0.5 0 0.2 0.4 0.6 0.8 1 ! 0 0 0.2 0.4 0.6 0.8 1 CS598JHM: Advanced NLP CS598JHM: Advanced NLP 15 16 ! Beta as prior for binomial So, what do we predict? Given a prior P(! |#,$) = Beta(#,$), and data D=(H,T), Our Bayesian estimate for the next coin flip P(x=1 | D): what is our posterior? 1 P (θ α, β,H,T) P (H, T θ)P (θ α, β) P (x = H D)= P (x = H θ)P (θ D)dθ | ∝ | | | 0 | | 1 H T α 1 β 1 = θP (θ D)dθ θ (1 θ) θ − (1 θ) − | ∝ − − 0 = E[θ D] | H+α 1 T +β 1 = E[Beta(H + α,T + β)] = θ − (1 θ) − − H + α = H + α + T + β With normalization Γ(H + α + T + β) H+α 1 T +β 1 P (θ α, β,H,T)= θ − (1 θ) − | Γ(H + α)Γ(T + β) − = Beta(α + H, β + T ) CS598JHM: Advanced NLP CS598JHM: Advanced NLP 17 18 Conjugate priors Multinomials: Dirichlet prior The beta distribution is a conjugate prior to the binomial: Multinomial distribution: the resulting posterior is also a beta distribution. Probability of observing each possible outcome ci exactly Xi times in a sequence of n yes/no trials: All members of the exponential family of distributions have n! N P (X =x ,...,X =x )= θx1 θxK if x = n conjugate priors. 1 i K k x ! x ! 1 ··· K i 1 K i=1 ··· Examples: - Multinomial: conjugate prior = Dirichlet Dirichlet prior: - Gaussian: conjugate prior = Gaussian Γ(α1 + ... + αk) αk 1 Dir(θ α1,...αk)= θk − | Γ(α1)...Γ(αk) k=1 CS598JHM: Advanced NLP CS598JHM: Advanced NLP 19 20 More about conjugate priors Todayʼs reading - We can interpret the hyperparameters as “pseudocounts” - Bishop, Pattern Recognition and Machine Learning, Ch. 2 - Sequential estimation (updating counts after each observation) gives same results as batch estimation - Add-one smoothing (Laplace smoothing) = uniform prior - On average, more data leads to a sharper posterior (sharper = lower variance) CS598JHM: Advanced NLP CS598JHM: Advanced NLP 21 22.
Recommended publications
  • The Beta-Binomial Distribution Introduction Bayesian Derivation

    The Beta-Binomial Distribution Introduction Bayesian Derivation

    In Danish: 2005-09-19 / SLB Translated: 2008-05-28 / SLB Bayesian Statistics, Simulation and Software The beta-binomial distribution I have translated this document, written for another course in Danish, almost as is. I have kept the references to Lee, the textbook used for that course. Introduction In Lee: Bayesian Statistics, the beta-binomial distribution is very shortly mentioned as the predictive distribution for the binomial distribution, given the conjugate prior distribution, the beta distribution. (In Lee, see pp. 78, 214, 156.) Here we shall treat it slightly more in depth, partly because it emerges in the WinBUGS example in Lee x 9.7, and partly because it possibly can be useful for your project work. Bayesian Derivation We make n independent Bernoulli trials (0-1 trials) with probability parameter π. It is well known that the number of successes x has the binomial distribution. Considering the probability parameter π unknown (but of course the sample-size parameter n is known), we have xjπ ∼ bin(n; π); where in generic1 notation n p(xjπ) = πx(1 − π)1−x; x = 0; 1; : : : ; n: x We assume as prior distribution for π a beta distribution, i.e. π ∼ beta(α; β); 1By this we mean that p is not a fixed function, but denotes the density function (in the discrete case also called the probability function) for the random variable the value of which is argument for p. 1 with density function 1 p(π) = πα−1(1 − π)β−1; 0 < π < 1: B(α; β) I remind you that the beta function can be expressed by the gamma function: Γ(α)Γ(β) B(α; β) = : (1) Γ(α + β) In Lee, x 3.1 is shown that the posterior distribution is a beta distribution as well, πjx ∼ beta(α + x; β + n − x): (Because of this result we say that the beta distribution is conjugate distribution to the binomial distribution.) We shall now derive the predictive distribution, that is finding p(x).
  • The Exponential Family 1 Definition

    The Exponential Family 1 Definition

    The Exponential Family David M. Blei Columbia University November 9, 2016 The exponential family is a class of densities (Brown, 1986). It encompasses many familiar forms of likelihoods, such as the Gaussian, Poisson, multinomial, and Bernoulli. It also encompasses their conjugate priors, such as the Gamma, Dirichlet, and beta. 1 Definition A probability density in the exponential family has this form p.x / h.x/ exp >t.x/ a./ ; (1) j D f g where is the natural parameter; t.x/ are sufficient statistics; h.x/ is the “base measure;” a./ is the log normalizer. Examples of exponential family distributions include Gaussian, gamma, Poisson, Bernoulli, multinomial, Markov models. Examples of distributions that are not in this family include student-t, mixtures, and hidden Markov models. (We are considering these families as distributions of data. The latent variables are implicitly marginalized out.) The statistic t.x/ is called sufficient because the probability as a function of only depends on x through t.x/. The exponential family has fundamental connections to the world of graphical models (Wainwright and Jordan, 2008). For our purposes, we’ll use exponential 1 families as components in directed graphical models, e.g., in the mixtures of Gaussians. The log normalizer ensures that the density integrates to 1, Z a./ log h.x/ exp >t.x/ d.x/ (2) D f g This is the negative logarithm of the normalizing constant. The function h.x/ can be a source of confusion. One way to interpret h.x/ is the (unnormalized) distribution of x when 0. It might involve statistics of x that D are not in t.x/, i.e., that do not vary with the natural parameter.
  • Machine Learning Conjugate Priors and Monte Carlo Methods

    Machine Learning Conjugate Priors and Monte Carlo Methods

    Hierarchical Bayes for Non-IID Data Conjugate Priors Monte Carlo Methods CPSC 540: Machine Learning Conjugate Priors and Monte Carlo Methods Mark Schmidt University of British Columbia Winter 2016 Hierarchical Bayes for Non-IID Data Conjugate Priors Monte Carlo Methods Admin Nothing exciting? We discussed empirical Bayes, where you optimize prior using marginal likelihood, Z argmax p(xjα; β) = argmax p(xjθ)p(θjα; β)dθ: α,β α,β θ Can be used to optimize λj, polynomial degree, RBF σi, polynomial vs. RBF, etc. We also considered hierarchical Bayes, where you put a prior on the prior, p(xjα; β)p(α; βjγ) p(α; βjx; γ) = : p(xjγ) But is the hyper-prior really needed? Hierarchical Bayes for Non-IID Data Conjugate Priors Monte Carlo Methods Last Time: Bayesian Statistics In Bayesian statistics we work with posterior over parameters, p(xjθ)p(θjα; β) p(θjx; α; β) = : p(xjα; β) We also considered hierarchical Bayes, where you put a prior on the prior, p(xjα; β)p(α; βjγ) p(α; βjx; γ) = : p(xjγ) But is the hyper-prior really needed? Hierarchical Bayes for Non-IID Data Conjugate Priors Monte Carlo Methods Last Time: Bayesian Statistics In Bayesian statistics we work with posterior over parameters, p(xjθ)p(θjα; β) p(θjx; α; β) = : p(xjα; β) We discussed empirical Bayes, where you optimize prior using marginal likelihood, Z argmax p(xjα; β) = argmax p(xjθ)p(θjα; β)dθ: α,β α,β θ Can be used to optimize λj, polynomial degree, RBF σi, polynomial vs.
  • 9. Binomial Distribution

    9. Binomial Distribution

    J. K. SHAH CLASSES Binomial Distribution 9. Binomial Distribution THEORETICAL DISTRIBUTION (Exists only in theory) 1. Theoretical Distribution is a distribution where the variables are distributed according to some definite mathematical laws. 2. In other words, Theoretical Distributions are mathematical models; where the frequencies / probabilities are calculated by mathematical computation. 3. Theoretical Distribution are also called as Expected Variance Distribution or Frequency Distribution 4. THEORETICAL DISTRIBUTION DISCRETE CONTINUOUS BINOMIAL POISSON Distribution Distribution NORMAL OR Student’s ‘t’ Chi-square Snedecor’s GAUSSIAN Distribution Distribution F-Distribution Distribution A. Binomial Distribution (Bernoulli Distribution) 1. This distribution is a discrete probability Distribution where the variable ‘x’ can assume only discrete values i.e. x = 0, 1, 2, 3,....... n 2. This distribution is derived from a special type of random experiment known as Bernoulli Experiment or Bernoulli Trials , which has the following characteristics (i) Each trial must be associated with two mutually exclusive & exhaustive outcomes – SUCCESS and FAILURE . Usually the probability of success is denoted by ‘p’ and that of the failure by ‘q’ where q = 1-p and therefore p + q = 1. (ii) The trials must be independent under identical conditions. (iii) The number of trial must be finite (countably finite). (iv) Probability of success and failure remains unchanged throughout the process. : 442 : J. K. SHAH CLASSES Binomial Distribution Note 1 : A ‘trial’ is an attempt to produce outcomes which is neither sure nor impossible in nature. Note 2 : The conditions mentioned may also be treated as the conditions for Binomial Distributions. 3. Characteristics or Properties of Binomial Distribution (i) It is a bi parametric distribution i.e.
  • Chapter 6 Continuous Random Variables and Probability

    Chapter 6 Continuous Random Variables and Probability

    EF 507 QUANTITATIVE METHODS FOR ECONOMICS AND FINANCE FALL 2019 Chapter 6 Continuous Random Variables and Probability Distributions Chap 6-1 Probability Distributions Probability Distributions Ch. 5 Discrete Continuous Ch. 6 Probability Probability Distributions Distributions Binomial Uniform Hypergeometric Normal Poisson Exponential Chap 6-2/62 Continuous Probability Distributions § A continuous random variable is a variable that can assume any value in an interval § thickness of an item § time required to complete a task § temperature of a solution § height in inches § These can potentially take on any value, depending only on the ability to measure accurately. Chap 6-3/62 Cumulative Distribution Function § The cumulative distribution function, F(x), for a continuous random variable X expresses the probability that X does not exceed the value of x F(x) = P(X £ x) § Let a and b be two possible values of X, with a < b. The probability that X lies between a and b is P(a < X < b) = F(b) -F(a) Chap 6-4/62 Probability Density Function The probability density function, f(x), of random variable X has the following properties: 1. f(x) > 0 for all values of x 2. The area under the probability density function f(x) over all values of the random variable X is equal to 1.0 3. The probability that X lies between two values is the area under the density function graph between the two values 4. The cumulative density function F(x0) is the area under the probability density function f(x) from the minimum x value up to x0 x0 f(x ) = f(x)dx 0 ò xm where
  • 3.2.3 Binomial Distribution

    3.2.3 Binomial Distribution

    3.2.3 Binomial Distribution The binomial distribution is based on the idea of a Bernoulli trial. A Bernoulli trail is an experiment with two, and only two, possible outcomes. A random variable X has a Bernoulli(p) distribution if 8 > <1 with probability p X = > :0 with probability 1 − p, where 0 ≤ p ≤ 1. The value X = 1 is often termed a “success” and X = 0 is termed a “failure”. The mean and variance of a Bernoulli(p) random variable are easily seen to be EX = (1)(p) + (0)(1 − p) = p and VarX = (1 − p)2p + (0 − p)2(1 − p) = p(1 − p). In a sequence of n identical, independent Bernoulli trials, each with success probability p, define the random variables X1,...,Xn by 8 > <1 with probability p X = i > :0 with probability 1 − p. The random variable Xn Y = Xi i=1 has the binomial distribution and it the number of sucesses among n independent trials. The probability mass function of Y is µ ¶ ¡ ¢ n ¡ ¢ P Y = y = py 1 − p n−y. y For this distribution, t n EX = np, Var(X) = np(1 − p),MX (t) = [pe + (1 − p)] . 1 Theorem 3.2.2 (Binomial theorem) For any real numbers x and y and integer n ≥ 0, µ ¶ Xn n (x + y)n = xiyn−i. i i=0 If we take x = p and y = 1 − p, we get µ ¶ Xn n 1 = (p + (1 − p))n = pi(1 − p)n−i. i i=0 Example 3.2.2 (Dice probabilities) Suppose we are interested in finding the probability of obtaining at least one 6 in four rolls of a fair die.
  • Solutions to the Exercises

    Solutions to the Exercises

    Solutions to the Exercises Chapter 1 Solution 1.1 (a) Your computer may be programmed to allocate borderline cases to the next group down, or the next group up; and it may or may not manage to follow this rule consistently, depending on its handling of the numbers involved. Following a rule which says 'move borderline cases to the next group up', these are the five classifications. (i) 1.0-1.2 1.2-1.4 1.4-1.6 1.6-1.8 1.8-2.0 2.0-2.2 2.2-2.4 6 6 4 8 4 3 4 2.4-2.6 2.6-2.8 2.8-3.0 3.0-3.2 3.2-3.4 3.4-3.6 3.6-3.8 6 3 2 2 0 1 1 (ii) 1.0-1.3 1.3-1.6 1.6-1.9 1.9-2.2 2.2-2.5 10 6 10 5 6 2.5-2.8 2.8-3.1 3.1-3.4 3.4-3.7 7 3 1 2 (iii) 0.8-1.1 1.1-1.4 1.4-1.7 1.7-2.0 2.0-2.3 2 10 6 10 7 2.3-2.6 2.6-2.9 2.9-3.2 3.2-3.5 3.5-3.8 6 4 3 1 1 (iv) 0.85-1.15 1.15-1.45 1.45-1.75 1.75-2.05 2.05-2.35 4 9 8 9 5 2.35-2.65 2.65-2.95 2.95-3.25 3.25-3.55 3.55-3.85 7 3 3 1 1 (V) 0.9-1.2 1.2-1.5 1.5-1.8 1.8-2.1 2.1-2.4 6 7 11 7 4 2.4-2.7 2.7-3.0 3.0-3.3 3.3-3.6 3.6-3.9 7 4 2 1 1 (b) Computer graphics: the diagrams are shown in Figures 1.9 to 1.11.
  • Chapter 5 Sections

    Chapter 5 Sections

    Chapter 5 Chapter 5 sections Discrete univariate distributions: 5.2 Bernoulli and Binomial distributions Just skim 5.3 Hypergeometric distributions 5.4 Poisson distributions Just skim 5.5 Negative Binomial distributions Continuous univariate distributions: 5.6 Normal distributions 5.7 Gamma distributions Just skim 5.8 Beta distributions Multivariate distributions Just skim 5.9 Multinomial distributions 5.10 Bivariate normal distributions 1 / 43 Chapter 5 5.1 Introduction Families of distributions How: Parameter and Parameter space pf /pdf and cdf - new notation: f (xj parameters ) Mean, variance and the m.g.f. (t) Features, connections to other distributions, approximation Reasoning behind a distribution Why: Natural justification for certain experiments A model for the uncertainty in an experiment All models are wrong, but some are useful – George Box 2 / 43 Chapter 5 5.2 Bernoulli and Binomial distributions Bernoulli distributions Def: Bernoulli distributions – Bernoulli(p) A r.v. X has the Bernoulli distribution with parameter p if P(X = 1) = p and P(X = 0) = 1 − p. The pf of X is px (1 − p)1−x for x = 0; 1 f (xjp) = 0 otherwise Parameter space: p 2 [0; 1] In an experiment with only two possible outcomes, “success” and “failure”, let X = number successes. Then X ∼ Bernoulli(p) where p is the probability of success. E(X) = p, Var(X) = p(1 − p) and (t) = E(etX ) = pet + (1 − p) 8 < 0 for x < 0 The cdf is F(xjp) = 1 − p for 0 ≤ x < 1 : 1 for x ≥ 1 3 / 43 Chapter 5 5.2 Bernoulli and Binomial distributions Binomial distributions Def: Binomial distributions – Binomial(n; p) A r.v.
  • Polynomial Singular Value Decompositions of a Family of Source-Channel Models

    Polynomial Singular Value Decompositions of a Family of Source-Channel Models

    Polynomial Singular Value Decompositions of a Family of Source-Channel Models The MIT Faculty has made this article openly available. Please share how this access benefits you. Your story matters. Citation Makur, Anuran and Lizhong Zheng. "Polynomial Singular Value Decompositions of a Family of Source-Channel Models." IEEE Transactions on Information Theory 63, 12 (December 2017): 7716 - 7728. © 2017 IEEE As Published http://dx.doi.org/10.1109/tit.2017.2760626 Publisher Institute of Electrical and Electronics Engineers (IEEE) Version Author's final manuscript Citable link https://hdl.handle.net/1721.1/131019 Terms of Use Creative Commons Attribution-Noncommercial-Share Alike Detailed Terms http://creativecommons.org/licenses/by-nc-sa/4.0/ IEEE TRANSACTIONS ON INFORMATION THEORY 1 Polynomial Singular Value Decompositions of a Family of Source-Channel Models Anuran Makur, Student Member, IEEE, and Lizhong Zheng, Fellow, IEEE Abstract—In this paper, we show that the conditional expec- interested in a particular subclass of one-parameter exponential tation operators corresponding to a family of source-channel families known as natural exponential families with quadratic models, defined by natural exponential families with quadratic variance functions (NEFQVF). So, we define natural exponen- variance functions and their conjugate priors, have orthonor- mal polynomials as singular vectors. These models include the tial families next. Gaussian channel with Gaussian source, the Poisson channel Definition 1 (Natural Exponential Family). Given a measur- with gamma source, and the binomial channel with beta source. To derive the singular vectors of these models, we prove and able space (Y; B(Y)) with a σ-finite measure µ, where Y ⊆ R employ the equivalent condition that their conditional moments and B(Y) denotes the Borel σ-algebra on Y, the parametrized are strictly degree preserving polynomials.
  • A Note on Skew-Normal Distribution Approximation to the Negative Binomal Distribution

    A Note on Skew-Normal Distribution Approximation to the Negative Binomal Distribution

    WSEAS TRANSACTIONS on MATHEMATICS Jyh-Jiuan Lin, Ching-Hui Chang, Rosemary Jou A NOTE ON SKEW-NORMAL DISTRIBUTION APPROXIMATION TO THE NEGATIVE BINOMAL DISTRIBUTION 1 2* 3 JYH-JIUAN LIN , CHING-HUI CHANG AND ROSEMARY JOU 1Department of Statistics Tamkang University 151 Ying-Chuan Road, Tamsui, Taipei County 251, Taiwan [email protected] 2Department of Applied Statistics and Information Science Ming Chuan University 5 De-Ming Road, Gui-Shan District, Taoyuan 333, Taiwan [email protected] 3 Graduate Institute of Management Sciences, Tamkang University 151 Ying-Chuan Road, Tamsui, Taipei County 251, Taiwan [email protected] Abstract: This article revisits the problem of approximating the negative binomial distribution. Its goal is to show that the skew-normal distribution, another alternative methodology, provides a much better approximation to negative binomial probabilities than the usual normal approximation. Key-Words: Central Limit Theorem, cumulative distribution function (cdf), skew-normal distribution, skew parameter. * Corresponding author ISSN: 1109-2769 32 Issue 1, Volume 9, January 2010 WSEAS TRANSACTIONS on MATHEMATICS Jyh-Jiuan Lin, Ching-Hui Chang, Rosemary Jou 1. INTRODUCTION where φ and Φ denote the standard normal pdf and cdf respectively. The This article revisits the problem of the parameter space is: μ ∈( −∞ , ∞), σ > 0 and negative binomial distribution approximation. Though the most commonly used solution of λ ∈(−∞ , ∞). Besides, the distribution approximating the negative binomial SN(0, 1, λ) with pdf distribution is by normal distribution, Peizer f( x | 0, 1, λ) = 2φ (x )Φ (λ x ) is called the and Pratt (1968) has investigated this problem standard skew-normal distribution.
  • A Compendium of Conjugate Priors

    A Compendium of Conjugate Priors

    A Compendium of Conjugate Priors Daniel Fink Environmental Statistics Group Department of Biology Montana State Univeristy Bozeman, MT 59717 May 1997 Abstract This report reviews conjugate priors and priors closed under sampling for a variety of data generating processes where the prior distributions are univariate, bivariate, and multivariate. The effects of transformations on conjugate prior relationships are considered and cases where conjugate prior relationships can be applied under transformations are identified. Univariate and bivariate prior relationships are verified using Monte Carlo methods. Contents 1 Introduction Experimenters are often in the position of having had collected some data from which they desire to make inferences about the process that produced that data. Bayes' theorem provides an appealing approach to solving such inference problems. Bayes theorem, π(θ) L(θ x ; : : : ; x ) g(θ x ; : : : ; x ) = j 1 n (1) j 1 n π(θ) L(θ x ; : : : ; x )dθ j 1 n is commonly interpreted in the following wayR. We want to make some sort of inference on the unknown parameter(s), θ, based on our prior knowledge of θ and the data collected, x1; : : : ; xn . Our prior knowledge is encapsulated by the probability distribution on θ; π(θ). The data that has been collected is combined with our prior through the likelihood function, L(θ x ; : : : ; x ) . The j 1 n normalized product of these two components yields a probability distribution of θ conditional on the data. This distribution, g(θ x ; : : : ; x ) , is known as the posterior distribution of θ. Bayes' j 1 n theorem is easily extended to cases where is θ multivariate, a vector of parameters.
  • Bayesian Filtering: from Kalman Filters to Particle Filters, and Beyond ZHE CHEN

    Bayesian Filtering: from Kalman Filters to Particle Filters, and Beyond ZHE CHEN

    MANUSCRIPT 1 Bayesian Filtering: From Kalman Filters to Particle Filters, and Beyond ZHE CHEN Abstract— In this self-contained survey/review paper, we system- IV Bayesian Optimal Filtering 9 atically investigate the roots of Bayesian filtering as well as its rich IV-AOptimalFiltering..................... 10 leaves in the literature. Stochastic filtering theory is briefly reviewed IV-BKalmanFiltering..................... 11 with emphasis on nonlinear and non-Gaussian filtering. Following IV-COptimumNonlinearFiltering.............. 13 the Bayesian statistics, different Bayesian filtering techniques are de- IV-C.1Finite-dimensionalFilters............ 13 veloped given different scenarios. Under linear quadratic Gaussian circumstance, the celebrated Kalman filter can be derived within the Bayesian framework. Optimal/suboptimal nonlinear filtering tech- V Numerical Approximation Methods 14 niques are extensively investigated. In particular, we focus our at- V-A Gaussian/Laplace Approximation ............ 14 tention on the Bayesian filtering approach based on sequential Monte V-BIterativeQuadrature................... 14 Carlo sampling, the so-called particle filters. Many variants of the V-C Mulitgrid Method and Point-Mass Approximation . 14 particle filter as well as their features (strengths and weaknesses) are V-D Moment Approximation ................. 15 discussed. Related theoretical and practical issues are addressed in V-E Gaussian Sum Approximation . ............. 16 detail. In addition, some other (new) directions on Bayesian filtering V-F Deterministic