DOCSLIB.ORG

A Class of Scale Mixtured Normal Distribution Rezaul Karim Md* Department of Statistics, Jahangirnagar University, Bangladesh

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
A Class of Scale Mixtured Normal Distribution Rezaul Karim Md* Department of Statistics, Jahangirnagar University, Bangladesh metrics & io B B i f o o s t l a a t i Rezaul Karim, J Biom Biostat 2016, 7:2 n s r t i u c o s J Journal of Biometrics & Biostatistics DOI: 10.4172/2155-6180.1000285 ISSN: 2155-6180 Research Article Article OpenOpen Access Access A Class of Scale Mixtured Normal Distribution Rezaul Karim Md* Department of Statistics, Jahangirnagar University, Bangladesh Abstract A class of scale mixtured normal distribution with lifetime probability distributions has been proposed in this article. Different moments, characteristic function and shape characteristics of these proposed probability distributions have also been provided. The scale mixture of normal distribution is extensively used in Biostatistical field. In this article, some new lifetime probability distributions have been proposed which is the scale mixture of normal distribution. Different properties such as moments, characteristic function, shape characteristics of these probability distributions have also been mentioned. Keyword: Scale mixture distribution; Inverted probability Main results distribution; Moments; Skewness; Kurtosis Here, the definition of scale mixture of normal distribution with Introduction well known lifetime distributions such as inverted chi-square, gamma, exponential, Rayleigh distributions are mentioned. The important A mixture distribution is a convex function of a set of probability characteristics of these distributions such as characteristic functions, distributions [1]. At first, Pearson considered a mixture distribution moments, and shape characteristics are also mentioned. The main of two normal distributions in 1894 and he studied the estimation results of the paper are presented in the form of definitions and of the parameters of his proposed distribution [2]. After around half theorems. century, some characteristics of mixture distributions were mentioned Definition 3.1 by Robins (1948). Some of researchers like Rider, Blichke, Karim [1,3-5] studied the various mixture of distributions. However, they A continuous random variable X is said to have a scale mixture of studied only for location mixture of the distributions though the scale normal distribution if its probability density function is given by 2 ∞ 1 x−µ mixture of the distributions have widely used in biostatistical field. The − 2 1 2σ location testing for Normal scale mixture distributions obtains in some fx(;µσ , )=∫ e h()τ d τ; −∞< x <∞ (3.1) 0 τσ2 π fastidious types of situations where the student’s t-test is inapplicable. Where, the parameters µ and σ2 satisfy − ∞ < µ < ∞ and σ2 > 0. Specifically, these cases permit tests of location to be prepared where Here h(τ) is a probability function of a random variable τ. The variable the assumption that sample observations obtain form populations X whose density function (3.1) is called a scale mixture normal variate having a Gaussian distribution can be replaced by assumption which can be written as X τ ~ N στµ 22 ),( that they arise from a Normal scale mixture distribution. A few researchers [3,6-8] mentioned the definition of scale mixture of normal Definition 3.2 distributions with its applications. There are many applications of scale If a random variable X has a probability density function mixture of normal distribution with lifetime probability distribution 11 = −∞< <∞ (3.2) in biostatistical field whereas; only the scale mixture of normal fx() n+1 ; x n 1 2 distribution is familiar for us. The scale mixture of normal distribution nσβ (,)x − µ 2 with lifetime distributions are not only used in biological field but also 22n σ used in fisheries, agricultural, psychological, paleontological field. It is also widely used in electrophoresis, and communication theories. In then it said to be scale mixture of normal distribution with inverted 2 this paper, the definition of the scale mixture of normal distribution chi-square distribution with location parameter µ, scale parameter σ with a set of lifetime probability distributions is defined initially and and shape parameter β. Various moments and shape characteristic of then scale mixture of normal distribution with inverted Chi-square, this probability distribution are presented by following theorem. Rayleigh, Exponential, Gamma and Erlang distributions are studied. Theorem 3.1 The different properties of these distributions are also presented. The moments and shape characteristics of scale mixture of normal Preliminaries A mixture distribution is the probability distribution of a random *Corresponding author: Rezaul Karim Md, Department of Statistics, Jahangirnagar variable whose values can be taken in a simple way from an underlying University, Bangladesh, Tel: 880-2-7791045-51; E-mail: [email protected] set of other random variables, with a certain probability of selection being associated with each. Suppose the random variable X has a Received February 27, 2016; Accepted March 01, 2016; Published March 08, 2016 probability density function (Pdf) f (x|θ). The sampling distribution of a statistic of the parameter θ is τ(θ)where Θ Citation: Rezaul Karim Md (2016) A Class of Scale Mixtured Normal Distribution. is the parameter space J Biom Biostat 7: 285. doi:10.4172/2155-6180.1000285 of θ. And then, a continuous mixture of densities f (x|θ) with weight functions τ(θ) is expressed by following: Copyright: © 2016 Rezaul Karim Md. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits mx()()()= ∫ f x|θτθ d θ. unrestricted use, distribution, and reproduction in any medium, provided the Θ original author and source are credited. J Biom Biostat ISSN: 2155-6180 JBMBS, an open access journal Volume 7 • Issue 2 • 1000285 Citation: Rezaul Karim Md (2016) A Class of Scale Mixtured Normal Distribution. J Biom Biostat 7: 285. doi:10.4172/2155-6180.1000285 Page 2 of 2 distribution with inverted chi-square distribution are given by Where µ, σ2 and β represent location, scale and shape parameters respectively. n 1 nrrσ 2 −+ rr 22 Theorem 3.4 µµ2rr= ;21+ = 0; n 1 The main characteristics of scale mixture normal distribution with 22 inverted exponential distribution are n 2 (2!r) Mean = 0, Variance = σ µ= σθ2rr1 −=r ; µ 0; n − 2 2rrr 21+ 32(n − ) 2!r Skewness: β = 0 , Kurtosis: β = . 2 1 2 n − 4 Mean=0 and variance = σ θ . Definition 3.3 Definition 3.6 A random variable X said to be scale mixture of normal distribution Let X be a continuous random variable. The probability density with inverted Gamma distribution if its probability density function is function of X is said to be a scale mixture normal distribution with following inverted Erlang distribution if its follows: nβ 1 = −∞< <∞ λθ fx() 1 ; x (3.3) 1 + = −∞ < < ∞ (3.6) 1 2 n fx() 1 ; x 2 θ + σ 2,Bnx − µ 1 2 2 21nβ + σ 2,Bnx − µ 2 σ 2 21λθ + σ Where, µ, σ2 and β represent location, scale and shape parameters respectively. Theorem 3.5 The different characteristics such as moments, skewnbess and Theorem 3.2 Kurtosis of the scale mixture normal distribution with inverted Erlang The main features such asr th central moment, Skewness, kurtosis, distribution are following: etc. of scale mixture of normal distribution with inverted Gamma θ − r distribution are following: (2!r) 2r r µ= σ( λθ) ; µ + = 0; Mean= 0, variance = 2,rm 2!r r λ 2r 1 (2!r) r nr− µ= σβ2r µ= 2rrr (n ) ;21+ 0; λθ 2!r n σ 2 θ −1 nβ 2 Mean= 0, variance = σ 3(θ −1) n −1 Skewness = 0, Kurtosis = . (θ − ) 3(n −1) 2 Skewness = 0, Kurtosis = . (n − 2) Concluding Remarks Definition 3.4 This article has been proposed a class of scale mixtured of normal distribution with different lifetime probability distribution such as The probability density function of a random variableX is said inverted Rayleigh distribution, Erlang distribution and explained to have a scale mixture of normal distribution with inverted Rayleigh some basic properties of these proposed probability distributions with distribution if its probability density function is defined by 1 applications in various fields of applied science and business. I believe, = −∞< <∞ fx() 3 ; x (3.4) this paper will help and encourage young researchers to stimulate 2 2 further research. 2 x − µ 2 σθ + 2 σ θ References Where θ is the parameter which satisfiesθ >0. 1. Karim MR, Hossain MP, Begum S, Hossain MF (2011) Rayleigh Mixture Distribution. Journal of Applied Mathematics. Theorem 3.3 2. Pearson K (1894) Contribution to the mathematical theory of evolution. The main properties of scale mixture normal distribution with Philosophical Transaction Royal Society. 185: 71-110. inverted Rayleigh distribution are given by following: 3. Andrews DF, Mallows CL (1974) Scale Mixtures of Normal Distribution. Journal 1− r of the Royal Statistical Society 36: 99-102. (2!r) 2r µ2rr= σµ;21+ = 0; 2!rrr θ 2 4. Rider PR (1961) The Method of Moments Applied To A Mixture of Two Exponential Distributions. The Annals of Mathematical Statistics 32: 143-147. σ 2 Mean=0 and variance = . θ 2 5. Blichke WR (1965) Mixtures of Discrete Distribution, Classical and Contagious Discrete Distributions. Calcutta Statistical Publishing Society Definition 3.5 Ed GP Patil 351- 372. The probability density function of scale mixture normal 6. Choy STB, Jennifer SKC (2008) Scale Mixtures Distributions in Statistical distribution with inverted exponential distribution is as follows Modelling. Australian & NewZealand Journal of Statistics 50: 135-146. θ = −∞< <∞ 7. West M (1987) On Scale Mixtures of Normal Distributions. Biometricka 74: fx() 3 , x (3.5) 2 646-648. x − µ 2 σθ+ 2 8. Monahan JF, Stefanski LA (1989) Normal scale mixture approximations to the σ logistic distribution with applications. Institute of Statistics Mineograph Series. J Biom Biostat ISSN: 2155-6180 JBMBS, an open access journal Volume 7 • Issue 2 • 1000285.
Load more

Recommended publications
  • A Tail Quantile Approximation Formula for the Student T and the Symmetric Generalized Hyperbolic Distribution
    A Service of Leibniz-Informationszentrum econstor Wirtschaft Leibniz Information Centre Make Your Publications Visible. zbw for Economics Schlüter, Stephan; Fischer, Matthias J. Working Paper A tail quantile approximation formula for the student t and the symmetric generalized hyperbolic distribution IWQW Discussion Papers, No. 05/2009 Provided in Cooperation with: Friedrich-Alexander University Erlangen-Nuremberg, Institute for Economics Suggested Citation: Schlüter, Stephan; Fischer, Matthias J. (2009) : A tail quantile approximation formula for the student t and the symmetric generalized hyperbolic distribution, IWQW Discussion Papers, No. 05/2009, Friedrich-Alexander-Universität Erlangen-Nürnberg, Institut für Wirtschaftspolitik und Quantitative Wirtschaftsforschung (IWQW), Nürnberg This Version is available at: http://hdl.handle.net/10419/29554 Standard-Nutzungsbedingungen: Terms of use: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Documents in EconStor may be saved and copied for your Zwecken und zum Privatgebrauch gespeichert und kopiert werden. personal and scholarly purposes. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle You are not to copy documents for public or commercial Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich purposes, to exhibit the documents publicly, to make them machen, vertreiben oder anderweitig nutzen. publicly available on the internet, or to distribute or otherwise use the documents in public. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, If the documents have been made available under an Open gelten abweichend von diesen Nutzungsbedingungen die in der dort Content Licence (especially Creative Commons Licences), you genannten Lizenz gewährten Nutzungsrechte. may exercise further usage rights as specified in the indicated licence. www.econstor.eu IWQW Institut für Wirtschaftspolitik und Quantitative Wirtschaftsforschung Diskussionspapier Discussion Papers No.
    [Show full text]
  • Mixture Random Effect Model Based Meta-Analysis for Medical Data
    Mixture Random Effect Model Based Meta-analysis For Medical Data Mining Yinglong Xia?, Shifeng Weng?, Changshui Zhang??, and Shao Li State Key Laboratory of Intelligent Technology and Systems,Department of Automation, Tsinghua University, Beijing, China [email protected], [email protected], fzcs,[email protected] Abstract. As a powerful tool for summarizing the distributed medi- cal information, Meta-analysis has played an important role in medical research in the past decades. In this paper, a more general statistical model for meta-analysis is proposed to integrate heterogeneous medi- cal researches efficiently. The novel model, named mixture random effect model (MREM), is constructed by Gaussian Mixture Model (GMM) and unifies the existing fixed effect model and random effect model. The pa- rameters of the proposed model are estimated by Markov Chain Monte Carlo (MCMC) method. Not only can MREM discover underlying struc- ture and intrinsic heterogeneity of meta datasets, but also can imply reasonable subgroup division. These merits embody the significance of our methods for heterogeneity assessment. Both simulation results and experiments on real medical datasets demonstrate the performance of the proposed model. 1 Introduction As the great improvement of experimental technologies, the growth of the vol- ume of scientific data relevant to medical experiment researches is getting more and more massively. However, often the results spreading over journals and on- line database appear inconsistent or even contradict because of variance of the studies. It makes the evaluation of those studies to be difficult. Meta-analysis is statistical technique for assembling to integrate the findings of a large collection of analysis results from individual studies.
    [Show full text]
  • A Family of Skew-Normal Distributions for Modeling Proportions and Rates with Zeros/Ones Excess
    S S symmetry Article A Family of Skew-Normal Distributions for Modeling Proportions and Rates with Zeros/Ones Excess Guillermo Martínez-Flórez 1, Víctor Leiva 2,* , Emilio Gómez-Déniz 3 and Carolina Marchant 4 1 Departamento de Matemáticas y Estadística, Facultad de Ciencias Básicas, Universidad de Córdoba, Montería 14014, Colombia; [email protected] 2 Escuela de Ingeniería Industrial, Pontificia Universidad Católica de Valparaíso, 2362807 Valparaíso, Chile 3 Facultad de Economía, Empresa y Turismo, Universidad de Las Palmas de Gran Canaria and TIDES Institute, 35001 Canarias, Spain; [email protected] 4 Facultad de Ciencias Básicas, Universidad Católica del Maule, 3466706 Talca, Chile; [email protected] * Correspondence: [email protected] or [email protected] Received: 30 June 2020; Accepted: 19 August 2020; Published: 1 September 2020 Abstract: In this paper, we consider skew-normal distributions for constructing new a distribution which allows us to model proportions and rates with zero/one inflation as an alternative to the inflated beta distributions. The new distribution is a mixture between a Bernoulli distribution for explaining the zero/one excess and a censored skew-normal distribution for the continuous variable. The maximum likelihood method is used for parameter estimation. Observed and expected Fisher information matrices are derived to conduct likelihood-based inference in this new type skew-normal distribution. Given the flexibility of the new distributions, we are able to show, in real data scenarios, the good performance of our proposal. Keywords: beta distribution; centered skew-normal distribution; maximum-likelihood methods; Monte Carlo simulations; proportions; R software; rates; zero/one inflated data 1.
    [Show full text]
  • D. Normal Mixture Models and Elliptical Models
    D. Normal Mixture Models and Elliptical Models 1. Normal Variance Mixtures 2. Normal Mean-Variance Mixtures 3. Spherical Distributions 4. Elliptical Distributions QRM 2010 74 D1. Multivariate Normal Mixture Distributions Pros of Multivariate Normal Distribution • inference is \well known" and estimation is \easy". • distribution is given by µ and Σ. • linear combinations are normal (! VaR and ES calcs easy). • conditional distributions are normal. > • For (X1;X2) ∼ N2(µ; Σ), ρ(X1;X2) = 0 () X1 and X2 are independent: QRM 2010 75 Multivariate Normal Variance Mixtures Cons of Multivariate Normal Distribution • tails are thin, meaning that extreme values are scarce in the normal model. • joint extremes in the multivariate model are also too scarce. • the distribution has a strong form of symmetry, called elliptical symmetry. How to repair the drawbacks of the multivariate normal model? QRM 2010 76 Multivariate Normal Variance Mixtures The random vector X has a (multivariate) normal variance mixture distribution if d p X = µ + WAZ; (1) where • Z ∼ Nk(0;Ik); • W ≥ 0 is a scalar random variable which is independent of Z; and • A 2 Rd×k and µ 2 Rd are a matrix and a vector of constants, respectively. > Set Σ := AA . Observe: XjW = w ∼ Nd(µ; wΣ). QRM 2010 77 Multivariate Normal Variance Mixtures Assumption: rank(A)= d ≤ k, so Σ is a positive definite matrix. If E(W ) < 1 then easy calculations give E(X) = µ and cov(X) = E(W )Σ: We call µ the location vector or mean vector and we call Σ the dispersion matrix. The correlation matrices of X and AZ are identical: corr(X) = corr(AZ): Multivariate normal variance mixtures provide the most useful examples of elliptical distributions.
    [Show full text]
  • Mixture Modeling with Applications in Alzheimer's Disease
    University of Kentucky UKnowledge Theses and Dissertations--Epidemiology and Biostatistics College of Public Health 2017 MIXTURE MODELING WITH APPLICATIONS IN ALZHEIMER'S DISEASE Frank Appiah University of Kentucky, [email protected] Digital Object Identifier: https://doi.org/10.13023/ETD.2017.100 Right click to open a feedback form in a new tab to let us know how this document benefits ou.y Recommended Citation Appiah, Frank, "MIXTURE MODELING WITH APPLICATIONS IN ALZHEIMER'S DISEASE" (2017). Theses and Dissertations--Epidemiology and Biostatistics. 14. https://uknowledge.uky.edu/epb_etds/14 This Doctoral Dissertation is brought to you for free and open access by the College of Public Health at UKnowledge. It has been accepted for inclusion in Theses and Dissertations--Epidemiology and Biostatistics by an authorized administrator of UKnowledge. For more information, please contact [email protected]. STUDENT AGREEMENT: I represent that my thesis or dissertation and abstract are my original work. Proper attribution has been given to all outside sources. I understand that I am solely responsible for obtaining any needed copyright permissions. I have obtained needed written permission statement(s) from the owner(s) of each third-party copyrighted matter to be included in my work, allowing electronic distribution (if such use is not permitted by the fair use doctrine) which will be submitted to UKnowledge as Additional File. I hereby grant to The University of Kentucky and its agents the irrevocable, non-exclusive, and royalty-free license to archive and make accessible my work in whole or in part in all forms of media, now or hereafter known.
    [Show full text]
  • A Two-Component Normal Mixture Alternative to the Fay-Herriot Model
    A two-component normal mixture alternative to the Fay-Herriot model August 22, 2015 Adrijo Chakraborty 1, Gauri Sankar Datta 2 3 4 and Abhyuday Mandal 5 Abstract: This article considers a robust hierarchical Bayesian approach to deal with ran- dom effects of small area means when some of these effects assume extreme values, resulting in outliers. In presence of outliers, the standard Fay-Herriot model, used for modeling area- level data, under normality assumptions of the random effects may overestimate random effects variance, thus provides less than ideal shrinkage towards the synthetic regression pre- dictions and inhibits borrowing information. Even a small number of substantive outliers of random effects result in a large estimate of the random effects variance in the Fay-Herriot model, thereby achieving little shrinkage to the synthetic part of the model or little reduction in posterior variance associated with the regular Bayes estimator for any of the small areas. While a scale mixture of normal distributions with known mixing distribution for the random effects has been found to be effective in presence of outliers, the solution depends on the mixing distribution. As a possible alternative solution to the problem, a two-component nor- mal mixture model has been proposed based on noninformative priors on the model variance parameters, regression coefficients and the mixing probability. Data analysis and simulation studies based on real, simulated and synthetic data show advantage of the proposed method over the standard Bayesian Fay-Herriot solution derived under normality of random effects. 1NORC at the University of Chicago, Bethesda, MD 20814, [email protected] 2Department of Statistics, University of Georgia, Athens, GA 30602, USA.
    [Show full text]
  • Student's T Distribution Based Estimation of Distribution
    1 Student’s t Distribution based Estimation of Distribution Algorithms for Derivative-free Global Optimization ∗Bin Liu Member, IEEE, Shi Cheng Member, IEEE, Yuhui Shi Fellow, IEEE Abstract In this paper, we are concerned with a branch of evolutionary algorithms termed estimation of distribution (EDA), which has been successfully used to tackle derivative-free global optimization problems. For existent EDA algorithms, it is a common practice to use a Gaussian distribution or a mixture of Gaussian components to represent the statistical property of available promising solutions found so far. Observing that the Student’s t distribution has heavier and longer tails than the Gaussian, which may be beneficial for exploring the solution space, we propose a novel EDA algorithm termed ESTDA, in which the Student’s t distribution, rather than Gaussian, is employed. To address hard multimodal and deceptive problems, we extend ESTDA further by substituting a single Student’s t distribution with a mixture of Student’s t distributions. The resulting algorithm is named as estimation of mixture of Student’s t distribution algorithm (EMSTDA). Both ESTDA and EMSTDA are evaluated through extensive and in-depth numerical experiments using over a dozen of benchmark objective functions. Empirical results demonstrate that the proposed algorithms provide remarkably better performance than their Gaussian counterparts. Index Terms derivative-free global optimization, EDA, estimation of distribution, Expectation-Maximization, mixture model, Student’s t distribution I. INTRODUCTION The estimation of distribution algorithm (EDA) is an evolutionary computation (EC) paradigm for tackling derivative-free global optimization problems [1]. EDAs have gained great success and attracted increasing attentions during the last decade arXiv:1608.03757v2 [cs.NE] 25 Nov 2016 [2].
    [Show full text]
  • Metaplus: Robust Meta-Analysis and Meta-Regression
    Package ‘metaplus’ June 12, 2021 Type Package Title Robust Meta-Analysis and Meta-Regression Version 1.0-2 Date 2021-06-10 Maintainer Ken Beath <[email protected]> Contact Ken Beath <[email protected]> Author Ken Beath [aut, cre], Ben Bolker [aut], R Development Core Team [aut] Description Performs meta-analysis and meta-regression using standard and robust methods with con- fidence intervals based on the profile likelihood. Robust methods are based on alternative distri- butions for the random effect, either the t-distribution (Lee and Thomp- son, 2008 <doi:10.1002/sim.2897> or Baker and Jackson, 2008 <doi:10.1007/s10729-007-9041- 8>) or mixtures of normals (Beath, 2014 <doi:10.1002/jrsm.1114>). Depends R(>= 3.2.0) Suggests R.rsp VignetteBuilder R.rsp Imports bbmle, metafor, boot, methods, numDeriv, MASS, graphics, stats, fastGHQuad, lme4, Rfast, parallel, doParallel, foreach, doRNG LazyData yes License GPL (>= 2) NeedsCompilation no Repository CRAN Date/Publication 2021-06-12 04:20:02 UTC R topics documented: metaplus-package . .2 AIC .............................................3 1 2 metaplus-package BIC .............................................3 cdp..............................................4 exercise . .5 logLik . .6 mag .............................................6 marinho . .7 metaplus . .8 outlierProbs . 12 plot.metaplus . 13 plot.outlierProbs . 14 summary . 15 testOutliers . 16 Index 18 metaplus-package Fits random effects meta-analysis models including robust models Description Allows fitting of random effects meta-analysis producing confidence intervals based on the profile likelihood (Hardy and Thompson, 1996). Two methods of robust meta-analysis are included, based on either the t-distribution (Baker and Jackson (2008) and Lee and Thompson (2008)) or normal- mixture distribution (Beath, 2014).
    [Show full text]
  • Bayesian Random-Effects Meta-Analysis Using the Bayesmeta
    JSS Journal of Statistical Software MMMMMM YYYY, Volume VV, Issue II. doi: 10.18637/jss.v000.i00 Bayesian random-effects meta-analysis using the bayesmeta R package Christian R¨over University Medical Center G¨ottingen Abstract The random-effects or normal-normal hierarchical model is commonly utilized in a wide range of meta-analysis applications. A Bayesian approach to inference is very at- tractive in this context, especially when a meta-analysis is based only on few studies. The bayesmeta R package provides readily accessible tools to perform Bayesian meta-analyses and generate plots and summaries, without having to worry about computational details. It allows for flexible prior specification and instant access to the resulting posterior distri- butions, including prediction and shrinkage estimation, and facilitating for example quick sensitivity checks. The present paper introduces the underlying theory and showcases its usage. Keywords: evidence synthesis, NNHM, between-study heterogeneity. 1. Introduction 1.1. Meta-analysis arXiv:1711.08683v2 [stat.CO] 10 Oct 2018 Evidence commonly comes in separate bits, and not necessarily from a single experiment. In contemporary science, the careful conduct of systematic reviews of the available evidence from diverse data sources is an effective and ubiquitously practiced means of compiling relevant in- formation. In this context, meta-analyses allow for the formal, mathematical combination of information to merge data from individual investigations to a joint result. Along with qualita- tive, often informal assessment and evaluation of the present evidence, meta-analytic methods have become a powerful tool to guide objective decision-making (Chalmers et al. 2002; Liberati et al.
    [Show full text]
  • Derivation of Mixture Distributions and Weighted Likelihood Function As Minimizers of Kl-Divergence Subject to Constraints
    Ann. Inst. Statist. Math. Vol. 57, No. 4, 687 701 (2005) @2005 The Institute of Statistical Mathematics DERIVATION OF MIXTURE DISTRIBUTIONS AND WEIGHTED LIKELIHOOD FUNCTION AS MINIMIZERS OF KL-DIVERGENCE SUBJECT TO CONSTRAINTS XIAOGANG WANG 1 AND JAMES V. ZIDEK 2 1Department of Mathematics and Statistics, York University, 4700 Keele Street, ON, Canada M3J 1P3 2Department of Statistics, University of British Columbia, 333-6356 Agriculture Road, BC, Canada V6T 1Z2 (Received May 26, 2003; revised July 7, 2004) Abstract. In this article, mixture distributions and weighted likelihoods are de- rived within an information-theoretic framework and shown to be closely related. This surprising relationship obtains in spite of the arithmetic form of the former and the geometric form of the latter. Mixture distributions are shown to be optima that minimize the entropy loss under certain constraints. The same framework implies the weighted likelihood when the distributions in the mixture are unknown and informa- tion from independent samples generated by them have to be used instead. Thus the likelihood weights trade bias for precision and yield inferential procedures such as estimates that can be more reliable than their classical counterparts. Key words and phrases: Euler-Lagrange equations, relative entropy, mixture distri- butions, weighted likelihood. 1. Introduction To introduce the results we present in this paper, suppose a statistician, S, is required to predict y, the realized value of a random variable Y with unknown probability density function (PDF), say f. Instead of a point prediction, S is allowed to state the prediction as a predictive PDF, say g. Finally, S receives a reward, log g(y).
    [Show full text]
  • Multivariate Distributions
    IEOR E4602: Quantitative Risk Management Spring 2016 c 2016 by Martin Haugh Multivariate Distributions We will study multivariate distributions in these notes, focusing1 in particular on multivariate normal, normal-mixture, spherical and elliptical distributions. In addition to studying their properties, we will also discuss techniques for simulating and, very briefly, estimating these distributions. Familiarity with these important classes of multivariate distributions is important for many aspects of risk management. We will defer the study of copulas until later in the course. 1 Preliminary Definitions Let X = (X1;:::Xn) be an n-dimensional vector of random variables. We have the following definitions and statements. > n Definition 1 (Joint CDF) For all x = (x1; : : : ; xn) 2 R , the joint cumulative distribution function (CDF) of X satisfies FX(x) = FX(x1; : : : ; xn) = P (X1 ≤ x1;:::;Xn ≤ xn): Definition 2 (Marginal CDF) For a fixed i, the marginal CDF of Xi satisfies FXi (xi) = FX(1;:::; 1; xi; 1;::: 1): It is straightforward to generalize the previous definition to joint marginal distributions. For example, the joint marginal distribution of Xi and Xj satisfies Fij(xi; xj) = FX(1;:::; 1; xi; 1;:::; 1; xj; 1;::: 1). If the joint CDF is absolutely continuous, then it has an associated probability density function (PDF) so that Z x1 Z xn FX(x1; : : : ; xn) = ··· f(u1; : : : ; un) du1 : : : dun: −∞ −∞ Similar statements also apply to the marginal CDF's. A collection of random variables is independent if the joint CDF (or PDF if it exists) can be factored into the product of the marginal CDFs (or PDFs). If > > X1 = (X1;:::;Xk) and X2 = (Xk+1;:::;Xn) is a partition of X then the conditional CDF satisfies FX2jX1 (x2jx1) = P (X2 ≤ x2jX1 = x1): If X has a PDF, f(·), then it satisfies Z xk+1 Z xn f(x1; : : : ; xk; uk+1; : : : ; un) FX2jX1 (x2jx1) = ··· duk+1 : : : dun −∞ −∞ fX1 (x1) where fX1 (·) is the joint marginal PDF of X1.
    [Show full text]
  • Fitting a Mixture Distribution to Data: Tutorial
    Fitting A Mixture Distribution to Data: Tutorial Benyamin Ghojogh [email protected] Department of Electrical and Computer Engineering, Machine Learning Laboratory, University of Waterloo, Waterloo, ON, Canada Aydin Ghojogh [email protected] Mark Crowley [email protected] Department of Electrical and Computer Engineering, Machine Learning Laboratory, University of Waterloo, Waterloo, ON, Canada Fakhri Karray [email protected] Department of Electrical and Computer Engineering, Centre for Pattern Analysis and Machine Intelligence, University of Waterloo, Waterloo, ON, Canada Abstract of K distributions fg1(x;Θ1); : : : ; gK (x;ΘK )g where the This paper is a step-by-step tutorial for fitting a weights fw1; : : : ; wK g sum to one. As is obvious, every mixture distribution to data. It merely assumes distribution in the mixture has its own parameter Θk. The the reader has the background of calculus and lin- mixture distribution is formulated as: ear algebra. Other required background is briefly K X reviewed before explaining the main algorithm. f(x;Θ1;:::; ΘK ) = wk gk(x;Θk); In explaining the main algorithm, first, fitting a k=1 (1) mixture of two distributions is detailed and ex- K X amples of fitting two Gaussians and Poissons, re- subject to wk = 1: spectively for continuous and discrete cases, are k=1 introduced. Thereafter, fitting several distribu- The distributions can be from different families, for ex- tions in general case is explained and examples ample from beta and normal distributions. However, this of several Gaussians (Gaussian Mixture Model) makes the problem very complex and sometimes useless; and Poissons are again provided. Model-based therefore, mostly the distributions in a mixture are from clustering, as one of the applications of mixture one family (e.g., all normal distributions) but with different distributions, is also introduced.
    [Show full text]