The Mixture Approach for Simulating New Families of Bivariate Distributions with Specified Correlations

Size: px
Start display at page:

Download "The Mixture Approach for Simulating New Families of Bivariate Distributions with Specified Correlations"

Transcription

1 The Mixture Approach for Simulating New Families of Bivariate Distributions with Specified Correlations John R. Michael, Significance, Inc. and William R. Schucany, Southern Methodist University The mixture approach is an exact methodology for simulating new families of bivariate distributions with specified correlation coefficients. It accommodates the entire range of correlations, produces bivariate surfaces that are intuitively appealing, and is often remarkably easy to implement. The approach is introduced in a Bayesian context and demonstrated for the conjugate families of beta and gamma distributions, with special attention given to the bivariate uniform. For these distributions, formulas for correlations have simple closed forms and computations are easy. KEY WORDS: Bayes, beta, conjugate prior, cool, exchangeable, gamma, generating, hierarchical models, Markov Chain Monte Carlo, posterior, uniform. 1. Introduction The generation of multivariate distributions has widespread applications for research and practice. An important application is the computer evaluation of new statistical methods for analyzing multivariate data. Many, if not most, multivariate methods assume a specific form for the underlying distribution of the observed data. Through simulation studies, the researcher can investigate properties of methods, estimators, or test statistics for a variety of multivariate distributions that may be encountered in practice. A popular multivariate distribution is the normal. It is easy to simulate, often permits closed-form theoretical results, and is perhaps the most widely known statistical distribution. The popularity of the normal might be due in part to its familiarity and convenience rather than to its appropriateness. Statistical distributions encountered in practice, particularly those underlying observational data, are often non-normal, e.g., skewed or relatively heavy-tailed. The development of convenient simulation techniques for alternative distributions can help increase their familiarity and use, both in research and practice. For non-normal data, some general approaches have been developed to

2 simulate multivariate distributions, such as the conditional distribution approach, the transformation approach, and the rejection approach. Unfortunately, these approaches often suffer from computational difficulties, are restrictive in form, and model only weak dependence (Johnson, 1987). This paper presents a new exact methodology for simulating multivariate distributions, which we term the mixture approach. It uses concepts that are found in Bayesian analysis, which inspired its development. Accordingly, to help fix ideas, we loosely refer to certain distributions in this paper as the prior, likelihood, joint, marginal, and posterior, where the likelihood is the conditional sampling distribution of the observed data. The essence of the mixture approach lies in a deceptively simple concept: for a specified prior, observations simulated from appropriate posteriors have the same marginal as the prior. The correlation between the two marginals depends on the intermediate likelihood for some disposable data that effectively "mixes" posteriors. Specifying the parameters of the likelihood precisely controls the correlation coefficient. In the language of Markov Chain Monte Carlo (MCMC) methods (see Tierney, 1994), we capitalize on the structure of sequences that are in equilibrium. In effect we exploit Gibbs samplers that have actually converged (see Casella and George, 1992). Section 2 presents the theory of the mixture approach. Section 3 summarizes the simulation steps using the mixture approach. Sections 4 and 5 apply the approach to create new bivariate beta and gamma families, respectively, both of which are remarkably easy to generate. Finally, Section 6 offers some concluding remarks and describes extensions that are underway. The appendices contain derivations of formulas for correlation coefficients and the density function for a specific bivariate uniform distribution. 2. Theoretical Basis of the Mixture Approach Let the random variable X 1 have a prior represented by the pdf g ( x 1 ; θ ), (1) where the parameter θ may be multidimensional. Next, conditioning on X 1 = x 1, let the random variable K for the data have the likelihood represented by the pmf (or pdf) h ( k x 1 ; η ), (2) where the parameter η may also be multidimensional. Multiplying (1) and (2) yields the joint of X 1 and K, 2

3 j ( x 1,k; θ,η ) = g ( x 1 ; θ) h ( k x 1 ; η ). (3) Integrating out x 1 yields the marginal of K, m ( k; θ,η ) = j ( z, k; θ, η ) dz. (4) Dividing (3) by (4) yields the posterior of X 2 given K = k, p ( x 2 k; θ,η ) = j ( x 2,k; θ,η ) / m ( k; θ, η ). (5) Formulas (1) through (4) parallel the conventional approach in Bayesian inference; however, the introduction of the new random variable X 2 in (5) is a fundamental departure. For continuous priors, the probability is zero that an observation, x 2, simulated from the posterior will exactly equal the corresponding value simulated from prior, x 1. Nevertheless, the marginal of X 2 is identical to the prior of X 1. This is because the simulation process makes no use of the data as a conditioning factor, thereby "disposing" of the intermediate information about K. The resulting distribution is a weighted average (mixture) of different posteriors that exactly reproduces the prior. We now formally demonstrate the equality of the marginal distributions of X 1 and X 2. Multiplying (3) and (5) yields the trivariate distribution of X 1, X 2, and K, f ( x 2,x 1,k; θ, η ) = j ( x 1, k; θ, η ) j ( x 2,k; θ, η ) / m ( k; θ, η ). (6) The symmetry of (6) with respect to x 1 and x 2 constitutes a proof that X 1 and X 2 have the same marginal distribution, g ( x 2 ; θ ). 3. Mixture Simulation Steps A bivariate pair ( x 1, x 2 ) is generated by sequentially simulating observations from the prior, likelihood, and posterior as follows: 1. Generate an observation, x 1, from the prior; 2. Generate an observation, k, from the likelihood, which is conditioned on x Generate an observation, x 2, from the posterior. Table 1 lists the steps in the mixture simulation approach alongside major steps in conventional Bayesian inference. 3

4 Table 1. Steps in Mixture Simulation and Conventional Bayesian Inference Step Mixture Simulation Bayesian Inference 1 Specify the prior. Specify the prior. 2 Simulate prior values, x 1. 3 Formulate likelihoods conditioned upon values x 1. Formulate the likelihood. 4 Simulate the data, k. from conditional likelihoods. Collect the data, k. 5 Derive conditional posteriors. Derive the posterior. 6 Simulate values, x 2, from conditional posteriors. 7 Form bivariate pairs ( x 1,x 2 ). Draw inferences about x Simulating a New Bivariate Beta Family The beta ( α, β ) is the familiar conjugate prior for estimating a binomial parameter x 1 in a Bayesian context. Applying the mixture approach, the associated prior, likelihood, posterior, and trivariate distributions are g ( x 1 ; α, β ) x 1 α-1 ( 1-x1 ) β-1, h ( k x 1 ; ν ) k C ν x 1 k ( 1-x1 ) ν-k, p ( x 2 k; α,β,ν ) x 2 α+k-1 ( 1-x2 ) β+ν k-1, and f ( x 1,x 2, k; α,β,ν ) ( x 1 x 2 ) α+k-1 [ ( 1-x 2 ) ( 1-x 1 ) ] β+ν-k-1, where θ = ( α, β ) in (1), η = ν in (2), and k C ν is the binomial coefficient. (Some proportionality constants have been omitted in the above formulas for the sake of simplicity.) The formula for the correlation coefficient, ρ beta, between X 1 and X 2 reduces to the simple intuitive expression ρ beta = ν / ( ν + α + β ). (7) 4

5 (See Appendix 1 for technical details.) Assuming the values of α and β are prespecified and fixed, the value of the correlation coefficient is controlled by specifying the value of ν, which represents the amount of information to be realized from the likelihood. Solving for ν, the required "effective" sample size for a specified correlation coefficient is ν = ( α + β ) ρ beta / ( 1-ρ beta ). (For the binomial case, if fractional sample sizes are needed, one can determine another member of the generalized power series family of distributions, which permits fractional values of ν, and can be used to expand the new bivariate beta family presented here. See Patil, Sharadchandra, & Rao, 1968, for details.) Inspection of (7) reveals that the entire range of correlations is accommodated by the mixture simulation approach. In particular, ρ beta is an increasing function of the effective sample size ( ν ) approaching 1 asymptotically, and ρ beta is a decreasing function of ( α + β ) approaching 0 asymptotically. Of course, negative correlations are obtained by reversing the sign of x 2. These properties parallel those in a Bayesian context: the dependence of the Bayes estimator on the data increases with the sample size from the likelihood, and decreases with the specificity of the prior. 4.1 Graphical Illustration of Mixture Simulation The following simple illustration is intended to elucidate the concepts upon which mixture simulation is based. We simulated five independent bivariate pairs of observations from the bivariate beta ( α = β = 3) and ρ beta = 0.75, which implies that ν = 2. Table 2 summarizes the relevant steps, which can be easily implemented, say, using the functions rbeta ( ) and rbinom ( ) in S-Plus 2000 (MathSoft, Inc., Seattle). It is instructive to follow the steps in the simulation of a single observation. The first x 1 value was 0.85, which produced k = 2, which led to x 2 = By disposing of our knowledge of which x 2 value was simulated from which posterior, we can now regard the sample of x 2 values as unconditioned and governed only by the prior. Thus the sample of x 2 values constitutes a random sample from a distribution that is precisely beta (3,3). The sample in Figure 1 has the observed correlation ρ beta = For illustrative purposes, we purposefully chose this sample from a number of simulated samples because of its regularity. Many of the other samples exhibited 5

6 substantially greater variability in some aspects. For example, in one simulation all five values of k were equal to 1. Table 2. Steps Used to Simulate Five Bivariate Pairs Step Operation Pair 1 Pair 2 Pair 3 Pair 4 Pair 5 1 Specified a beta (3, 3) prior, graphed in Figure 1a. 2 Simulated five independent values of X 1, marked in Figure 1a. 3 Simulated five values of K from conditional binomials (2,x 1 ), marked in Figure 1b. 4 Formed three distinct posteriors, overlaid in Figure 1c, and Simulated five values of X 2, marked in Figure 1c. 5 Formed ( x 1, x 2 ) pairs, plotted in Figure 1d (5,3) (4,4) (3,5) (5,3) (4,4)

7 (a) Beta(3, 3) Prior Density (b)conditional Binomial Data (c) Three Beta Posteriors g(x1) p(x k) x1 (d) Scatterplot for Bivariate Beta k x2 x2 x1 Figure 1. Illustrative Simulation of a New Bivariate Beta (3,3) 7

8 Howard (1998) argues that it may be appropriate to use dependent priors for a Bayesian analysis of two binomials. One of his main recommendations is that Bayesians give careful consideration to dependent priors. Our bivariate beta provides a very satisfactory alternative to the augmented versions of independent betas presented by Howard (1998). Furthermore, the mixture approach has the advantage of preserving the marginals. The Dirichlet distribution is a special three-parameter bivariate beta family that has been previously investigated. Loukas (1984) examines several alternative approaches to generating bivariate beta variates and recommends combining independent gammas for reasons of efficiency. This family is not flexible enough to permit a range of correlations. In this regard see Gupta and Wong (1985) for an extension of the Morgenstern system. 4.2 Importance of the Bivariate Uniform When α = β = 1, the beta reverts to the familiar uniform on the unit interval. Efficient simulation of correlated uniform observations is particularly useful because bivariate distributions with other specified marginals can then be obtained by applying inverse probability integral transformations. In particular, suppose we wish to simulate ( Y, Z), wherey and Z have marginal cdf's F and G and continuous quantile functions (inverse marginal cdf's) V ( u ) = F -1 ( u ) and W ( u ) = G -1 ( u ), respectively. It follows that ( Y, Z ) = [V ( X 1 ), W ( X 2 ) ] has the desired bivariate distribution. 8

9 rho = 0.5 rho = 0.7 x2 x2 x1 x1 rho = 0.9 rho = 0.95 x2 x2 x1 x1 Figure 2. Scatterplots of 100 Observations Simulated from the Bivariate Beta (3, 3) with Correlations = 0.5, 0.7, 0.9,

10 For some distributions, the appropriate quantile function can be applied with relative ease. For example, if Y = V ( X 1 ) = µ + σ { log [X 1 / (1-X 1 ) ] ) }, then Y is logistic with location and scale parameters µ and σ, respectively. Although the correlation between simulated bivariate uniforms can be precisely controlled with the mixture approach, the correlation between the transformed bivariate uniforms, Y and Z, depends on the functional form of the transformation and must be derived on a case-by-case basis. A Taylor series expansion of the transformation function might provide a satisfactory approximation. Conversely, any continuous bivariate distribution can be transformed to bivariate uniformity by applying the appropriate probability integral transformation to each member of the pair. In particular, (U 1,U 2 ) = [F (Y ), G (Z ) ] is bivariate uniform. This distribution is generally not a member of the new bivariate family of uniforms generated by the mixture approach; however, a member of the latter family might provide a satisfactory approximation to the distribution of (U 1,U 2 ). Assuming so, it is then possible to simulate the bivariate pair (Y,Z ) by applying the appropriate quantile functions, as described in the previous paragraph. In practice, when the data are observational, the functional forms of the true cdf's are typically unknown and can be unique to the process that generates the data. In such cases, smoothed empirical cdf's and corresponding quantile functions might be adequate for transforming the observed data to and from approximate uniformity, thereby permitting simulation of quite specialized joint distributions. 4.3 Graphical Illustration of the Bivariate Uniform Figure 3 is a contour plot of a bivariate uniform generated by the mixture method. The special case where ν = 2 and ρ beta = 0.5 is plotted because of the relative simplicity of the joint distribution of X 1 and X 2, which has pdf b ( x 1, x 2 ) = 3{ x 1 ( x 1-2 ) + x 2 ( x 2-2 ) x 1 x x 1 x 2 ( 1- x 1 ) ( 1- x 2 ) }. (See Appendix 2 for technical details.) 10

11 The contour plot in Figure 3 bears some resemblance to the corresponding plot for the Farlie-Gumbel-Morgenstern family of bivariate distributions (see Johnson, 1987, page 181). However the latter family is subject to the restriction that ρ beta 1 / 3, whereas the mixture approach accommodates the entire range of correlation coefficients. Falk (1999) describes a distinctly different approach using probability integral transformations of multivariate normals. Although Falk s approach allows a prescribed correlation structure, there are some constraints for a positive semi-definite result. x x1 Figure 3. Contour Plot of the Bivariate Uniform Distribution with r beta =

12 5. Simulating a New Bivariate Gamma Family The gamma ( λ ) is the familiar conjugate prior for estimating a Poisson parameter x 1 in a Bayesian context. The presentation here parallels that for the beta in the previous section. Applying the mixture approach, the associated prior, likelihood, posterior, and trivariate distributions are g ( x 1 ; λ ) x 1 r-1 e -λx 1, h ( k x 1 ; τ ) x 1 k e -τx 1 / k!, p ( x 2 k; λ,τ ) x 2 r+k-1 e -(λ+τ) x 2, and f ( x 2, k, x 1 ; λ,τ ) ( x 1 x 2 ) r+k-1 e -(λ+τ) (x 2 +x 1 ), where θ = ( r,λ ) and η = τ. Once again, the formula for the correlation coefficient between X 1 and X 2 reduces to a simple intuitive expression, ρ gamma = τ / ( τ + λ ). (8) (See Appendix 3 for technical details). Assuming the value of λ is pre-specified and fixed, the value of the correlation coefficient is controlled by specifying the value of τ, which represents the amount of the data observed from the likelihood. Solving for τ, the required Poisson parameter for a specified correlation coefficient is τ = λρ gamma / ( 1-ρ gamma ). 6. Concluding Remarks The mixture approach is a method of simulating new families of bivariate distributions. The approach has the advantage of accommodating the entire range of correlation coefficients, which can be precisely controlled. And for some distributions, such as the beta and gamma, the computations are very easy. Application of the mixture approach with a conjugate normal prior for a normal mean yields the familiar bivariate normal distribution. Although this approach is new, it is not very efficient, compared to the standard simulation approach (see Johnson, 1987, pages 52-54), because it requires the generation of three independent normals rather than two. 12

13 Manuscripts are in preparation that extend the mixture approach with specified correlations in two directions: to the multivariate case with three or more random variables; and to the bivariate beta with different (arbitrary) marginal distributions. The latter extension is particularly relevant to problems encountered in practice where marginal distributions have different shapes. References Casella, G., & George, E.I. (1992). Explaining the Gibbs Sampler, The American Statistician, 46, Falk, M. (1999). A simple approach to the generation of uniformly distributed random variables with prescribed correlations. Communications in Statistics - Simulation and Computation, 28, Gupta, A.K. & Wong, C.F. (1985). On three and five parameter bivariate beta distributions, Metrika, 32, Howard, J. V. (1998). The 2x2 table: A discussion from a Bayesian viewpoint, Statistical Science, 13, Johnson, M.E. (1987). Multivariate Statistical Simulation. New York: John Wiley & Sons. Loukas, S. (1984). Simple methods for computer generation of bivariate beta random variables, J. Statistical Computation and Simulation, 20, Patil, G. P., Sharadchandra, W. J., & Rao, C. R. (1968). A Dictionary and Bibliography of Discrete Distributions. New York: Hafner. Tierney, L. (1994). Markov Chains for exploring posterior distributions (with discussion), The Annals of Statistics, 22,

14

15

16

17

18

Computational statistics

Computational statistics Computational statistics Markov Chain Monte Carlo methods Thierry Denœux March 2017 Thierry Denœux Computational statistics March 2017 1 / 71 Contents of this chapter When a target density f can be evaluated

More information

Bayesian Regression Linear and Logistic Regression

Bayesian Regression Linear and Logistic Regression When we want more than point estimates Bayesian Regression Linear and Logistic Regression Nicole Beckage Ordinary Least Squares Regression and Lasso Regression return only point estimates But what if we

More information

Motivation Scale Mixutres of Normals Finite Gaussian Mixtures Skew-Normal Models. Mixture Models. Econ 690. Purdue University

Motivation Scale Mixutres of Normals Finite Gaussian Mixtures Skew-Normal Models. Mixture Models. Econ 690. Purdue University Econ 690 Purdue University In virtually all of the previous lectures, our models have made use of normality assumptions. From a computational point of view, the reason for this assumption is clear: combined

More information

Markov Chain Monte Carlo methods

Markov Chain Monte Carlo methods Markov Chain Monte Carlo methods Tomas McKelvey and Lennart Svensson Signal Processing Group Department of Signals and Systems Chalmers University of Technology, Sweden November 26, 2012 Today s learning

More information

Bayesian Methods for Machine Learning

Bayesian Methods for Machine Learning Bayesian Methods for Machine Learning CS 584: Big Data Analytics Material adapted from Radford Neal s tutorial (http://ftp.cs.utoronto.ca/pub/radford/bayes-tut.pdf), Zoubin Ghahramni (http://hunch.net/~coms-4771/zoubin_ghahramani_bayesian_learning.pdf),

More information

2 Inference for Multinomial Distribution

2 Inference for Multinomial Distribution Markov Chain Monte Carlo Methods Part III: Statistical Concepts By K.B.Athreya, Mohan Delampady and T.Krishnan 1 Introduction In parts I and II of this series it was shown how Markov chain Monte Carlo

More information

eqr094: Hierarchical MCMC for Bayesian System Reliability

eqr094: Hierarchical MCMC for Bayesian System Reliability eqr094: Hierarchical MCMC for Bayesian System Reliability Alyson G. Wilson Statistical Sciences Group, Los Alamos National Laboratory P.O. Box 1663, MS F600 Los Alamos, NM 87545 USA Phone: 505-667-9167

More information

ABC methods for phase-type distributions with applications in insurance risk problems

ABC methods for phase-type distributions with applications in insurance risk problems ABC methods for phase-type with applications problems Concepcion Ausin, Department of Statistics, Universidad Carlos III de Madrid Joint work with: Pedro Galeano, Universidad Carlos III de Madrid Simon

More information

10. Exchangeability and hierarchical models Objective. Recommended reading

10. Exchangeability and hierarchical models Objective. Recommended reading 10. Exchangeability and hierarchical models Objective Introduce exchangeability and its relation to Bayesian hierarchical models. Show how to fit such models using fully and empirical Bayesian methods.

More information

13: Variational inference II

13: Variational inference II 10-708: Probabilistic Graphical Models, Spring 2015 13: Variational inference II Lecturer: Eric P. Xing Scribes: Ronghuo Zheng, Zhiting Hu, Yuntian Deng 1 Introduction We started to talk about variational

More information

Stat 451 Lecture Notes Markov Chain Monte Carlo. Ryan Martin UIC

Stat 451 Lecture Notes Markov Chain Monte Carlo. Ryan Martin UIC Stat 451 Lecture Notes 07 12 Markov Chain Monte Carlo Ryan Martin UIC www.math.uic.edu/~rgmartin 1 Based on Chapters 8 9 in Givens & Hoeting, Chapters 25 27 in Lange 2 Updated: April 4, 2016 1 / 42 Outline

More information

Markov Chain Monte Carlo (MCMC)

Markov Chain Monte Carlo (MCMC) Markov Chain Monte Carlo (MCMC Dependent Sampling Suppose we wish to sample from a density π, and we can evaluate π as a function but have no means to directly generate a sample. Rejection sampling can

More information

Bayesian Inference in GLMs. Frequentists typically base inferences on MLEs, asymptotic confidence

Bayesian Inference in GLMs. Frequentists typically base inferences on MLEs, asymptotic confidence Bayesian Inference in GLMs Frequentists typically base inferences on MLEs, asymptotic confidence limits, and log-likelihood ratio tests Bayesians base inferences on the posterior distribution of the unknowns

More information

Hyperparameter estimation in Dirichlet process mixture models

Hyperparameter estimation in Dirichlet process mixture models Hyperparameter estimation in Dirichlet process mixture models By MIKE WEST Institute of Statistics and Decision Sciences Duke University, Durham NC 27706, USA. SUMMARY In Bayesian density estimation and

More information

David B. Dahl. Department of Statistics, and Department of Biostatistics & Medical Informatics University of Wisconsin Madison

David B. Dahl. Department of Statistics, and Department of Biostatistics & Medical Informatics University of Wisconsin Madison AN IMPROVED MERGE-SPLIT SAMPLER FOR CONJUGATE DIRICHLET PROCESS MIXTURE MODELS David B. Dahl dbdahl@stat.wisc.edu Department of Statistics, and Department of Biostatistics & Medical Informatics University

More information

Stat 451 Lecture Notes Simulating Random Variables

Stat 451 Lecture Notes Simulating Random Variables Stat 451 Lecture Notes 05 12 Simulating Random Variables Ryan Martin UIC www.math.uic.edu/~rgmartin 1 Based on Chapter 6 in Givens & Hoeting, Chapter 22 in Lange, and Chapter 2 in Robert & Casella 2 Updated:

More information

CPSC 540: Machine Learning

CPSC 540: Machine Learning CPSC 540: Machine Learning MCMC and Non-Parametric Bayes Mark Schmidt University of British Columbia Winter 2016 Admin I went through project proposals: Some of you got a message on Piazza. No news is

More information

Principles of Bayesian Inference

Principles of Bayesian Inference Principles of Bayesian Inference Sudipto Banerjee University of Minnesota July 20th, 2008 1 Bayesian Principles Classical statistics: model parameters are fixed and unknown. A Bayesian thinks of parameters

More information

Fall 2017 STAT 532 Homework Peter Hoff. 1. Let P be a probability measure on a collection of sets A.

Fall 2017 STAT 532 Homework Peter Hoff. 1. Let P be a probability measure on a collection of sets A. 1. Let P be a probability measure on a collection of sets A. (a) For each n N, let H n be a set in A such that H n H n+1. Show that P (H n ) monotonically converges to P ( k=1 H k) as n. (b) For each n

More information

MCMC and Gibbs Sampling. Sargur Srihari

MCMC and Gibbs Sampling. Sargur Srihari MCMC and Gibbs Sampling Sargur srihari@cedar.buffalo.edu 1 Topics 1. Markov Chain Monte Carlo 2. Markov Chains 3. Gibbs Sampling 4. Basic Metropolis Algorithm 5. Metropolis-Hastings Algorithm 6. Slice

More information

Multivariate Distribution Models

Multivariate Distribution Models Multivariate Distribution Models Model Description While the probability distribution for an individual random variable is called marginal, the probability distribution for multiple random variables is

More information

Bayesian Methods with Monte Carlo Markov Chains II

Bayesian Methods with Monte Carlo Markov Chains II Bayesian Methods with Monte Carlo Markov Chains II Henry Horng-Shing Lu Institute of Statistics National Chiao Tung University hslu@stat.nctu.edu.tw http://tigpbp.iis.sinica.edu.tw/courses.htm 1 Part 3

More information

Introduction to Markov Chain Monte Carlo & Gibbs Sampling

Introduction to Markov Chain Monte Carlo & Gibbs Sampling Introduction to Markov Chain Monte Carlo & Gibbs Sampling Prof. Nicholas Zabaras Sibley School of Mechanical and Aerospace Engineering 101 Frank H. T. Rhodes Hall Ithaca, NY 14853-3801 Email: zabaras@cornell.edu

More information

Bayesian nonparametric estimation of finite population quantities in absence of design information on nonsampled units

Bayesian nonparametric estimation of finite population quantities in absence of design information on nonsampled units Bayesian nonparametric estimation of finite population quantities in absence of design information on nonsampled units Sahar Z Zangeneh Robert W. Keener Roderick J.A. Little Abstract In Probability proportional

More information

Contents. Part I: Fundamentals of Bayesian Inference 1

Contents. Part I: Fundamentals of Bayesian Inference 1 Contents Preface xiii Part I: Fundamentals of Bayesian Inference 1 1 Probability and inference 3 1.1 The three steps of Bayesian data analysis 3 1.2 General notation for statistical inference 4 1.3 Bayesian

More information

Ronald Christensen. University of New Mexico. Albuquerque, New Mexico. Wesley Johnson. University of California, Irvine. Irvine, California

Ronald Christensen. University of New Mexico. Albuquerque, New Mexico. Wesley Johnson. University of California, Irvine. Irvine, California Texts in Statistical Science Bayesian Ideas and Data Analysis An Introduction for Scientists and Statisticians Ronald Christensen University of New Mexico Albuquerque, New Mexico Wesley Johnson University

More information

A Search and Jump Algorithm for Markov Chain Monte Carlo Sampling. Christopher Jennison. Adriana Ibrahim. Seminar at University of Kuwait

A Search and Jump Algorithm for Markov Chain Monte Carlo Sampling. Christopher Jennison. Adriana Ibrahim. Seminar at University of Kuwait A Search and Jump Algorithm for Markov Chain Monte Carlo Sampling Christopher Jennison Department of Mathematical Sciences, University of Bath, UK http://people.bath.ac.uk/mascj Adriana Ibrahim Institute

More information

STA 4273H: Statistical Machine Learning

STA 4273H: Statistical Machine Learning STA 4273H: Statistical Machine Learning Russ Salakhutdinov Department of Computer Science! Department of Statistical Sciences! rsalakhu@cs.toronto.edu! h0p://www.cs.utoronto.ca/~rsalakhu/ Lecture 7 Approximate

More information

F denotes cumulative density. denotes probability density function; (.)

F denotes cumulative density. denotes probability density function; (.) BAYESIAN ANALYSIS: FOREWORDS Notation. System means the real thing and a model is an assumed mathematical form for the system.. he probability model class M contains the set of the all admissible models

More information

COS513 LECTURE 8 STATISTICAL CONCEPTS

COS513 LECTURE 8 STATISTICAL CONCEPTS COS513 LECTURE 8 STATISTICAL CONCEPTS NIKOLAI SLAVOV AND ANKUR PARIKH 1. MAKING MEANINGFUL STATEMENTS FROM JOINT PROBABILITY DISTRIBUTIONS. A graphical model (GM) represents a family of probability distributions

More information

Metropolis Hastings. Rebecca C. Steorts Bayesian Methods and Modern Statistics: STA 360/601. Module 9

Metropolis Hastings. Rebecca C. Steorts Bayesian Methods and Modern Statistics: STA 360/601. Module 9 Metropolis Hastings Rebecca C. Steorts Bayesian Methods and Modern Statistics: STA 360/601 Module 9 1 The Metropolis-Hastings algorithm is a general term for a family of Markov chain simulation methods

More information

Stat 5101 Lecture Notes

Stat 5101 Lecture Notes Stat 5101 Lecture Notes Charles J. Geyer Copyright 1998, 1999, 2000, 2001 by Charles J. Geyer May 7, 2001 ii Stat 5101 (Geyer) Course Notes Contents 1 Random Variables and Change of Variables 1 1.1 Random

More information

Bayesian inference. Fredrik Ronquist and Peter Beerli. October 3, 2007

Bayesian inference. Fredrik Ronquist and Peter Beerli. October 3, 2007 Bayesian inference Fredrik Ronquist and Peter Beerli October 3, 2007 1 Introduction The last few decades has seen a growing interest in Bayesian inference, an alternative approach to statistical inference.

More information

Dynamic System Identification using HDMR-Bayesian Technique

Dynamic System Identification using HDMR-Bayesian Technique Dynamic System Identification using HDMR-Bayesian Technique *Shereena O A 1) and Dr. B N Rao 2) 1), 2) Department of Civil Engineering, IIT Madras, Chennai 600036, Tamil Nadu, India 1) ce14d020@smail.iitm.ac.in

More information

HANDBOOK OF APPLICABLE MATHEMATICS

HANDBOOK OF APPLICABLE MATHEMATICS HANDBOOK OF APPLICABLE MATHEMATICS Chief Editor: Walter Ledermann Volume VI: Statistics PART A Edited by Emlyn Lloyd University of Lancaster A Wiley-Interscience Publication JOHN WILEY & SONS Chichester

More information

Markov Chain Monte Carlo

Markov Chain Monte Carlo Markov Chain Monte Carlo Recall: To compute the expectation E ( h(y ) ) we use the approximation E(h(Y )) 1 n n h(y ) t=1 with Y (1),..., Y (n) h(y). Thus our aim is to sample Y (1),..., Y (n) from f(y).

More information

Default Priors and Effcient Posterior Computation in Bayesian

Default Priors and Effcient Posterior Computation in Bayesian Default Priors and Effcient Posterior Computation in Bayesian Factor Analysis January 16, 2010 Presented by Eric Wang, Duke University Background and Motivation A Brief Review of Parameter Expansion Literature

More information

Bayesian philosophy Bayesian computation Bayesian software. Bayesian Statistics. Petter Mostad. Chalmers. April 6, 2017

Bayesian philosophy Bayesian computation Bayesian software. Bayesian Statistics. Petter Mostad. Chalmers. April 6, 2017 Chalmers April 6, 2017 Bayesian philosophy Bayesian philosophy Bayesian statistics versus classical statistics: War or co-existence? Classical statistics: Models have variables and parameters; these are

More information

Slice Sampling with Adaptive Multivariate Steps: The Shrinking-Rank Method

Slice Sampling with Adaptive Multivariate Steps: The Shrinking-Rank Method Slice Sampling with Adaptive Multivariate Steps: The Shrinking-Rank Method Madeleine B. Thompson Radford M. Neal Abstract The shrinking rank method is a variation of slice sampling that is efficient at

More information

Bayesian Life Test Planning for the Weibull Distribution with Given Shape Parameter

Bayesian Life Test Planning for the Weibull Distribution with Given Shape Parameter Statistics Preprints Statistics 10-8-2002 Bayesian Life Test Planning for the Weibull Distribution with Given Shape Parameter Yao Zhang Iowa State University William Q. Meeker Iowa State University, wqmeeker@iastate.edu

More information

Computation and Monte Carlo Techniques

Computation and Monte Carlo Techniques Computation and Monte Carlo Techniques Up to now we have seen conjugate Bayesian analysis: posterior prior likelihood beta(a + x, b + n x) beta(a, b) binomial(x; n) θ a+x 1 (1 θ) b+n x 1 θ a 1 (1 θ) b

More information

David Giles Bayesian Econometrics

David Giles Bayesian Econometrics David Giles Bayesian Econometrics 5. Bayesian Computation Historically, the computational "cost" of Bayesian methods greatly limited their application. For instance, by Bayes' Theorem: p(θ y) = p(θ)p(y

More information

Bayesian Inference for Discretely Sampled Diffusion Processes: A New MCMC Based Approach to Inference

Bayesian Inference for Discretely Sampled Diffusion Processes: A New MCMC Based Approach to Inference Bayesian Inference for Discretely Sampled Diffusion Processes: A New MCMC Based Approach to Inference Osnat Stramer 1 and Matthew Bognar 1 Department of Statistics and Actuarial Science, University of

More information

Control Variates for Markov Chain Monte Carlo

Control Variates for Markov Chain Monte Carlo Control Variates for Markov Chain Monte Carlo Dellaportas, P., Kontoyiannis, I., and Tsourti, Z. Dept of Statistics, AUEB Dept of Informatics, AUEB 1st Greek Stochastics Meeting Monte Carlo: Probability

More information

Bayesian inference for multivariate extreme value distributions

Bayesian inference for multivariate extreme value distributions Bayesian inference for multivariate extreme value distributions Sebastian Engelke Clément Dombry, Marco Oesting Toronto, Fields Institute, May 4th, 2016 Main motivation For a parametric model Z F θ of

More information

STAT 518 Intro Student Presentation

STAT 518 Intro Student Presentation STAT 518 Intro Student Presentation Wen Wei Loh April 11, 2013 Title of paper Radford M. Neal [1999] Bayesian Statistics, 6: 475-501, 1999 What the paper is about Regression and Classification Flexible

More information

Parameter Estimation. William H. Jefferys University of Texas at Austin Parameter Estimation 7/26/05 1

Parameter Estimation. William H. Jefferys University of Texas at Austin Parameter Estimation 7/26/05 1 Parameter Estimation William H. Jefferys University of Texas at Austin bill@bayesrules.net Parameter Estimation 7/26/05 1 Elements of Inference Inference problems contain two indispensable elements: Data

More information

Appendix F. Computational Statistics Toolbox. The Computational Statistics Toolbox can be downloaded from:

Appendix F. Computational Statistics Toolbox. The Computational Statistics Toolbox can be downloaded from: Appendix F Computational Statistics Toolbox The Computational Statistics Toolbox can be downloaded from: http://www.infinityassociates.com http://lib.stat.cmu.edu. Please review the readme file for installation

More information

On Reparametrization and the Gibbs Sampler

On Reparametrization and the Gibbs Sampler On Reparametrization and the Gibbs Sampler Jorge Carlos Román Department of Mathematics Vanderbilt University James P. Hobert Department of Statistics University of Florida March 2014 Brett Presnell Department

More information

Bayesian Inference. Chapter 2: Conjugate models

Bayesian Inference. Chapter 2: Conjugate models Bayesian Inference Chapter 2: Conjugate models Conchi Ausín and Mike Wiper Department of Statistics Universidad Carlos III de Madrid Master in Business Administration and Quantitative Methods Master in

More information

Stat 451 Lecture Notes Monte Carlo Integration

Stat 451 Lecture Notes Monte Carlo Integration Stat 451 Lecture Notes 06 12 Monte Carlo Integration Ryan Martin UIC www.math.uic.edu/~rgmartin 1 Based on Chapter 6 in Givens & Hoeting, Chapter 23 in Lange, and Chapters 3 4 in Robert & Casella 2 Updated:

More information

Bayesian Statistical Methods. Jeff Gill. Department of Political Science, University of Florida

Bayesian Statistical Methods. Jeff Gill. Department of Political Science, University of Florida Bayesian Statistical Methods Jeff Gill Department of Political Science, University of Florida 234 Anderson Hall, PO Box 117325, Gainesville, FL 32611-7325 Voice: 352-392-0262x272, Fax: 352-392-8127, Email:

More information

Markov Chain Monte Carlo Methods

Markov Chain Monte Carlo Methods Markov Chain Monte Carlo Methods John Geweke University of Iowa, USA 2005 Institute on Computational Economics University of Chicago - Argonne National Laboaratories July 22, 2005 The problem p (θ, ω I)

More information

MCMC Methods: Gibbs and Metropolis

MCMC Methods: Gibbs and Metropolis MCMC Methods: Gibbs and Metropolis Patrick Breheny February 28 Patrick Breheny BST 701: Bayesian Modeling in Biostatistics 1/30 Introduction As we have seen, the ability to sample from the posterior distribution

More information

Part III. A Decision-Theoretic Approach and Bayesian testing

Part III. A Decision-Theoretic Approach and Bayesian testing Part III A Decision-Theoretic Approach and Bayesian testing 1 Chapter 10 Bayesian Inference as a Decision Problem The decision-theoretic framework starts with the following situation. We would like to

More information

CHAPTER 1 INTRODUCTION

CHAPTER 1 INTRODUCTION CHAPTER 1 INTRODUCTION 1.0 Discrete distributions in statistical analysis Discrete models play an extremely important role in probability theory and statistics for modeling count data. The use of discrete

More information

Metropolis-Hastings Algorithm

Metropolis-Hastings Algorithm Strength of the Gibbs sampler Metropolis-Hastings Algorithm Easy algorithm to think about. Exploits the factorization properties of the joint probability distribution. No difficult choices to be made to

More information

DAG models and Markov Chain Monte Carlo methods a short overview

DAG models and Markov Chain Monte Carlo methods a short overview DAG models and Markov Chain Monte Carlo methods a short overview Søren Højsgaard Institute of Genetics and Biotechnology University of Aarhus August 18, 2008 Printed: August 18, 2008 File: DAGMC-Lecture.tex

More information

Linear Models A linear model is defined by the expression

Linear Models A linear model is defined by the expression Linear Models A linear model is defined by the expression x = F β + ɛ. where x = (x 1, x 2,..., x n ) is vector of size n usually known as the response vector. β = (β 1, β 2,..., β p ) is the transpose

More information

A Bayesian perspective on GMM and IV

A Bayesian perspective on GMM and IV A Bayesian perspective on GMM and IV Christopher A. Sims Princeton University sims@princeton.edu November 26, 2013 What is a Bayesian perspective? A Bayesian perspective on scientific reporting views all

More information

The Bayesian Choice. Christian P. Robert. From Decision-Theoretic Foundations to Computational Implementation. Second Edition.

The Bayesian Choice. Christian P. Robert. From Decision-Theoretic Foundations to Computational Implementation. Second Edition. Christian P. Robert The Bayesian Choice From Decision-Theoretic Foundations to Computational Implementation Second Edition With 23 Illustrations ^Springer" Contents Preface to the Second Edition Preface

More information

36-463/663: Hierarchical Linear Models

36-463/663: Hierarchical Linear Models 36-463/663: Hierarchical Linear Models Taste of MCMC / Bayes for 3 or more levels Brian Junker 132E Baker Hall brian@stat.cmu.edu 1 Outline Practical Bayes Mastery Learning Example A brief taste of JAGS

More information

Approximate Bayesian computation for spatial extremes via open-faced sandwich adjustment

Approximate Bayesian computation for spatial extremes via open-faced sandwich adjustment Approximate Bayesian computation for spatial extremes via open-faced sandwich adjustment Ben Shaby SAMSI August 3, 2010 Ben Shaby (SAMSI) OFS adjustment August 3, 2010 1 / 29 Outline 1 Introduction 2 Spatial

More information

Hastings-within-Gibbs Algorithm: Introduction and Application on Hierarchical Model

Hastings-within-Gibbs Algorithm: Introduction and Application on Hierarchical Model UNIVERSITY OF TEXAS AT SAN ANTONIO Hastings-within-Gibbs Algorithm: Introduction and Application on Hierarchical Model Liang Jing April 2010 1 1 ABSTRACT In this paper, common MCMC algorithms are introduced

More information

FAV i R This paper is produced mechanically as part of FAViR. See for more information.

FAV i R This paper is produced mechanically as part of FAViR. See  for more information. Bayesian Claim Severity Part 2 Mixed Exponentials with Trend, Censoring, and Truncation By Benedict Escoto FAV i R This paper is produced mechanically as part of FAViR. See http://www.favir.net for more

More information

Estimation of Quantiles

Estimation of Quantiles 9 Estimation of Quantiles The notion of quantiles was introduced in Section 3.2: recall that a quantile x α for an r.v. X is a constant such that P(X x α )=1 α. (9.1) In this chapter we examine quantiles

More information

Multivariate Normal & Wishart

Multivariate Normal & Wishart Multivariate Normal & Wishart Hoff Chapter 7 October 21, 2010 Reading Comprehesion Example Twenty-two children are given a reading comprehsion test before and after receiving a particular instruction method.

More information

Subject CS1 Actuarial Statistics 1 Core Principles

Subject CS1 Actuarial Statistics 1 Core Principles Institute of Actuaries of India Subject CS1 Actuarial Statistics 1 Core Principles For 2019 Examinations Aim The aim of the Actuarial Statistics 1 subject is to provide a grounding in mathematical and

More information

Lecture 16: Mixtures of Generalized Linear Models

Lecture 16: Mixtures of Generalized Linear Models Lecture 16: Mixtures of Generalized Linear Models October 26, 2006 Setting Outline Often, a single GLM may be insufficiently flexible to characterize the data Setting Often, a single GLM may be insufficiently

More information

Bayesian Inference and MCMC

Bayesian Inference and MCMC Bayesian Inference and MCMC Aryan Arbabi Partly based on MCMC slides from CSC412 Fall 2018 1 / 18 Bayesian Inference - Motivation Consider we have a data set D = {x 1,..., x n }. E.g each x i can be the

More information

Lecture 13 : Variational Inference: Mean Field Approximation

Lecture 13 : Variational Inference: Mean Field Approximation 10-708: Probabilistic Graphical Models 10-708, Spring 2017 Lecture 13 : Variational Inference: Mean Field Approximation Lecturer: Willie Neiswanger Scribes: Xupeng Tong, Minxing Liu 1 Problem Setup 1.1

More information

STA 216, GLM, Lecture 16. October 29, 2007

STA 216, GLM, Lecture 16. October 29, 2007 STA 216, GLM, Lecture 16 October 29, 2007 Efficient Posterior Computation in Factor Models Underlying Normal Models Generalized Latent Trait Models Formulation Genetic Epidemiology Illustration Structural

More information

Bayesian nonparametric models for bipartite graphs

Bayesian nonparametric models for bipartite graphs Bayesian nonparametric models for bipartite graphs François Caron Department of Statistics, Oxford Statistics Colloquium, Harvard University November 11, 2013 F. Caron 1 / 27 Bipartite networks Readers/Customers

More information

Bayesian Graphical Models

Bayesian Graphical Models Graphical Models and Inference, Lecture 16, Michaelmas Term 2009 December 4, 2009 Parameter θ, data X = x, likelihood L(θ x) p(x θ). Express knowledge about θ through prior distribution π on θ. Inference

More information

Principles of Bayesian Inference

Principles of Bayesian Inference Principles of Bayesian Inference Sudipto Banerjee and Andrew O. Finley 2 Biostatistics, School of Public Health, University of Minnesota, Minneapolis, Minnesota, U.S.A. 2 Department of Forestry & Department

More information

On prediction and density estimation Peter McCullagh University of Chicago December 2004

On prediction and density estimation Peter McCullagh University of Chicago December 2004 On prediction and density estimation Peter McCullagh University of Chicago December 2004 Summary Having observed the initial segment of a random sequence, subsequent values may be predicted by calculating

More information

Monte Carlo in Bayesian Statistics

Monte Carlo in Bayesian Statistics Monte Carlo in Bayesian Statistics Matthew Thomas SAMBa - University of Bath m.l.thomas@bath.ac.uk December 4, 2014 Matthew Thomas (SAMBa) Monte Carlo in Bayesian Statistics December 4, 2014 1 / 16 Overview

More information

Statistical Inference

Statistical Inference Statistical Inference Robert L. Wolpert Institute of Statistics and Decision Sciences Duke University, Durham, NC, USA Spring, 2006 1. DeGroot 1973 In (DeGroot 1973), Morrie DeGroot considers testing the

More information

Robust estimation of skew-normal distribution with location and scale parameters via log-regularly varying functions

Robust estimation of skew-normal distribution with location and scale parameters via log-regularly varying functions Robust estimation of skew-normal distribution with location and scale parameters via log-regularly varying functions Shintaro Hashimoto Department of Mathematics, Hiroshima University October 5, 207 Abstract

More information

Bayesian Nonparametrics

Bayesian Nonparametrics Bayesian Nonparametrics Peter Orbanz Columbia University PARAMETERS AND PATTERNS Parameters P(X θ) = Probability[data pattern] 3 2 1 0 1 2 3 5 0 5 Inference idea data = underlying pattern + independent

More information

Bayesian Nonparametric Models

Bayesian Nonparametric Models Bayesian Nonparametric Models David M. Blei Columbia University December 15, 2015 Introduction We have been looking at models that posit latent structure in high dimensional data. We use the posterior

More information

Principles of Bayesian Inference

Principles of Bayesian Inference Principles of Bayesian Inference Sudipto Banerjee 1 and Andrew O. Finley 2 1 Biostatistics, School of Public Health, University of Minnesota, Minneapolis, Minnesota, U.S.A. 2 Department of Forestry & Department

More information

Likelihood-free MCMC

Likelihood-free MCMC Bayesian inference for stable distributions with applications in finance Department of Mathematics University of Leicester September 2, 2011 MSc project final presentation Outline 1 2 3 4 Classical Monte

More information

Markov Switching Regular Vine Copulas

Markov Switching Regular Vine Copulas Int. Statistical Inst.: Proc. 58th World Statistical Congress, 2011, Dublin (Session CPS057) p.5304 Markov Switching Regular Vine Copulas Stöber, Jakob and Czado, Claudia Lehrstuhl für Mathematische Statistik,

More information

Hierarchical models. Dr. Jarad Niemi. August 31, Iowa State University. Jarad Niemi (Iowa State) Hierarchical models August 31, / 31

Hierarchical models. Dr. Jarad Niemi. August 31, Iowa State University. Jarad Niemi (Iowa State) Hierarchical models August 31, / 31 Hierarchical models Dr. Jarad Niemi Iowa State University August 31, 2017 Jarad Niemi (Iowa State) Hierarchical models August 31, 2017 1 / 31 Normal hierarchical model Let Y ig N(θ g, σ 2 ) for i = 1,...,

More information

MCMC algorithms for fitting Bayesian models

MCMC algorithms for fitting Bayesian models MCMC algorithms for fitting Bayesian models p. 1/1 MCMC algorithms for fitting Bayesian models Sudipto Banerjee sudiptob@biostat.umn.edu University of Minnesota MCMC algorithms for fitting Bayesian models

More information

Exponential Families

Exponential Families Exponential Families David M. Blei 1 Introduction We discuss the exponential family, a very flexible family of distributions. Most distributions that you have heard of are in the exponential family. Bernoulli,

More information

CSC 2541: Bayesian Methods for Machine Learning

CSC 2541: Bayesian Methods for Machine Learning CSC 2541: Bayesian Methods for Machine Learning Radford M. Neal, University of Toronto, 2011 Lecture 10 Alternatives to Monte Carlo Computation Since about 1990, Markov chain Monte Carlo has been the dominant

More information

Estimation of Copula Models with Discrete Margins (via Bayesian Data Augmentation) Michael S. Smith

Estimation of Copula Models with Discrete Margins (via Bayesian Data Augmentation) Michael S. Smith Estimation of Copula Models with Discrete Margins (via Bayesian Data Augmentation) Michael S. Smith Melbourne Business School, University of Melbourne (Joint with Mohamad Khaled, University of Queensland)

More information

Latent Variable Models for Binary Data. Suppose that for a given vector of explanatory variables x, the latent

Latent Variable Models for Binary Data. Suppose that for a given vector of explanatory variables x, the latent Latent Variable Models for Binary Data Suppose that for a given vector of explanatory variables x, the latent variable, U, has a continuous cumulative distribution function F (u; x) and that the binary

More information

How to select a good vine

How to select a good vine Universitetet i Oslo ingrihaf@math.uio.no International FocuStat Workshop on Focused Information Criteria and Related Themes, May 9-11, 2016 Copulae Regular vines Model selection and reduction Limitations

More information

PROBABILITY DISTRIBUTIONS. J. Elder CSE 6390/PSYC 6225 Computational Modeling of Visual Perception

PROBABILITY DISTRIBUTIONS. J. Elder CSE 6390/PSYC 6225 Computational Modeling of Visual Perception PROBABILITY DISTRIBUTIONS Credits 2 These slides were sourced and/or modified from: Christopher Bishop, Microsoft UK Parametric Distributions 3 Basic building blocks: Need to determine given Representation:

More information

Probability and Estimation. Alan Moses

Probability and Estimation. Alan Moses Probability and Estimation Alan Moses Random variables and probability A random variable is like a variable in algebra (e.g., y=e x ), but where at least part of the variability is taken to be stochastic.

More information

Gentle Introduction to Infinite Gaussian Mixture Modeling

Gentle Introduction to Infinite Gaussian Mixture Modeling Gentle Introduction to Infinite Gaussian Mixture Modeling with an application in neuroscience By Frank Wood Rasmussen, NIPS 1999 Neuroscience Application: Spike Sorting Important in neuroscience and for

More information

Doing Bayesian Integrals

Doing Bayesian Integrals ASTR509-13 Doing Bayesian Integrals The Reverend Thomas Bayes (c.1702 1761) Philosopher, theologian, mathematician Presbyterian (non-conformist) minister Tunbridge Wells, UK Elected FRS, perhaps due to

More information

STA 4273H: Statistical Machine Learning

STA 4273H: Statistical Machine Learning STA 4273H: Statistical Machine Learning Russ Salakhutdinov Department of Statistics! rsalakhu@utstat.toronto.edu! http://www.utstat.utoronto.ca/~rsalakhu/ Sidney Smith Hall, Room 6002 Lecture 3 Linear

More information

Stat 535 C - Statistical Computing & Monte Carlo Methods. Arnaud Doucet.

Stat 535 C - Statistical Computing & Monte Carlo Methods. Arnaud Doucet. Stat 535 C - Statistical Computing & Monte Carlo Methods Arnaud Doucet Email: arnaud@cs.ubc.ca 1 1.1 Outline Introduction to Markov chain Monte Carlo The Gibbs Sampler Examples Overview of the Lecture

More information

Web Appendix for Hierarchical Adaptive Regression Kernels for Regression with Functional Predictors by D. B. Woodard, C. Crainiceanu, and D.

Web Appendix for Hierarchical Adaptive Regression Kernels for Regression with Functional Predictors by D. B. Woodard, C. Crainiceanu, and D. Web Appendix for Hierarchical Adaptive Regression Kernels for Regression with Functional Predictors by D. B. Woodard, C. Crainiceanu, and D. Ruppert A. EMPIRICAL ESTIMATE OF THE KERNEL MIXTURE Here we

More information

HANDBOOK OF APPLICABLE MATHEMATICS

HANDBOOK OF APPLICABLE MATHEMATICS HANDBOOK OF APPLICABLE MATHEMATICS Chief Editor: Walter Ledermann Volume II: Probability Emlyn Lloyd University oflancaster A Wiley-Interscience Publication JOHN WILEY & SONS Chichester - New York - Brisbane

More information

CSC 2541: Bayesian Methods for Machine Learning

CSC 2541: Bayesian Methods for Machine Learning CSC 2541: Bayesian Methods for Machine Learning Radford M. Neal, University of Toronto, 2011 Lecture 4 Problem: Density Estimation We have observed data, y 1,..., y n, drawn independently from some unknown

More information