DOCSLIB.ORG

Generalized Linear Models & Logistic Regression

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Generalized Linear Models & Logistic Regression Generalized Linear Models & Logistic Regression Jonathan Pillow Mathematical Tools for Neuroscience (NEU 314) Spring, 2016 lecture 17 Example 3: unknown neuron 100 75 50 (spike count) 25 0 -25 0 25 (contrast) What model would you use to fit this neuron? Linear-Nonlinear-Poisson model stimulus filter nonlinearity Poisson spiking λ(t) stimulus k f conditional intensity (“spike rate”) Poisson spiking • example of generalized linear model (GLM) Aside on GLMs: 1. Be careful about terminology: GLM ≠ GLM General Linear Model Generalized Linear Model (Nelder 1972) Linear Linear 2003 interview with John Nelder... Stephen Senn: I must confess to having some confusion when I was a young statistician between general linear models and generalized linear models. Do you regret the terminology? John Nelder: I think probably I do. I suspect we should have found some more fancy name for it that would have stuck and not been confused with the general linear model, although general and generalized are not quite the same. I can see why it might have been better to have thought of something else. Senn, (2003). Statistical Science Moral: Be careful when naming your model! 2. General Linear Model Linear Noise (exponential family) “Dimensionality Reduction” Examples: 1. Gaussian 2. Poisson 3. Generalized Linear Model Linear Nonlinear Noise (exponential family) Examples: 1. Gaussian 2. Poisson Linear-Nonlinear-Poisson stimulus filter exponential Poisson nonlinearity spiking stimulus k f λ(t) conditional intensity (“spike rate”) • output: Poisson process (Truccolo et al 04) GLM with spike-history dependence stimulus filter exponential probabilistic nonlinearity spiking stimulus + post-spike filter conditional intensity (spike rate) • output: product of stimulus and spike-history term GLM dynamic behaviors • irregular spiking (Poisson process) stimulus filter outputs (“currents”) p(spike) post-spike filter h(t) GLM dynamic behaviors • regular spiking stimulus filter outputs (“currents”) p(spike) post-spike filter h(t) GLM dynamic behaviors • bursting stimulus filter outputs (“currents”) p(spike) post-spike filter h(t) GLM dynamic behaviors • adaptation stimulus filter outputs (“currents”) p(spike) post-spike filter h(t) multi-neuron GLM stimulus filter exponential probabilistic nonlinearity spiking + post-spike filter neuron 1 stimulus + neuron 2 multi-neuron GLM stimulus filter exponential probabilistic nonlinearity spiking + post-spike filter neuron 1 stimulus coupling filters + neuron 2 GLM equivalent diagram: ... time t conditional intensity (spike rate) Logistic Regression GLM for binary classification weights sigmoid Bernoulli input “0” or “1” 1. Linear weights 2. sigmoid (“logistic”) function 3. Bernoulli (coin flip) Logistic Regression GLM for binary classification weights sigmoid Bernoulli input “0” or “1” compact expression: (note when y = 1, this is equal to exp(wx)/(1+exp(wx)), which is equal to 1/(1+exp(-wx) ) Logistic Regression GLM for binary classification weights sigmoid Bernoulli input “0” or “1” compact expression: fit w by maximizing log-likelihood: Logistic Regression geometric view p = 0.5 contour (classification boundary) neuron 2spikes neuron 1 spikes Bayesian Estimation three basic ingredients: 1. Likelihood jointly determine the posterior 2. Prior “cost” of making an estimate 3. Loss function L(✓ˆ, ✓) if the true value is • fully specifies how to generate an estimate from the data Bayesian estimator is defined as: ✓ˆ(m) = arg min L(✓ˆ, ✓)p(✓ m)d✓ “Bayes’ risk” ˆ | ✓ Z Typical Loss functions and Bayesian estimators 2 1. L(✓ˆ, ✓)=(✓ˆ ✓) squared error loss − 0 need to find minimizing the expected loss: Differentiate with respect to and set to zero: “posterior mean” also known as Bayes’ Least Squares (BLS) estimator Typical Loss functions and Bayesian estimators ˆ ˆ “zero-one” loss 2. L(✓, ✓)=1 δ(✓ ✓) (1 unless ) − − 0 expected loss: which is minimized by: • posterior maximum (or “mode”). • known as maximum a posteriori (MAP) estimate..
Load more

Recommended publications
  • A Generalized Linear Model for Principal Component Analysis of Binary Data
    A Generalized Linear Model for Principal Component Analysis of Binary Data Andrew I. Schein Lawrence K. Saul Lyle H. Ungar Department of Computer and Information Science University of Pennsylvania Moore School Building 200 South 33rd Street Philadelphia, PA 19104-6389 {ais,lsaul,ungar}@cis.upenn.edu Abstract they are not generally appropriate for other data types. Recently, Collins et al.[5] derived generalized criteria We investigate a generalized linear model for for dimensionality reduction by appealing to proper- dimensionality reduction of binary data. The ties of distributions in the exponential family. In their model is related to principal component anal- framework, the conventional PCA of real-valued data ysis (PCA) in the same way that logistic re- emerges naturally from assuming a Gaussian distribu- gression is related to linear regression. Thus tion over a set of observations, while generalized ver- we refer to the model as logistic PCA. In this sions of PCA for binary and nonnegative data emerge paper, we derive an alternating least squares respectively by substituting the Bernoulli and Pois- method to estimate the basis vectors and gen- son distributions for the Gaussian. For binary data, eralized linear coefficients of the logistic PCA the generalized model's relationship to PCA is anal- model. The resulting updates have a simple ogous to the relationship between logistic and linear closed form and are guaranteed at each iter- regression[12]. In particular, the model exploits the ation to improve the model's likelihood. We log-odds as the natural parameter of the Bernoulli dis- evaluate the performance of logistic PCA|as tribution and the logistic function as its canonical link.
    [Show full text]
  • Generalized Linear Models (Glms)
    San Jos´eState University Math 261A: Regression Theory & Methods Generalized Linear Models (GLMs) Dr. Guangliang Chen This lecture is based on the following textbook sections: • Chapter 13: 13.1 – 13.3 Outline of this presentation: • What is a GLM? • Logistic regression • Poisson regression Generalized Linear Models (GLMs) What is a GLM? In ordinary linear regression, we assume that the response is a linear function of the regressors plus Gaussian noise: 0 2 y = β0 + β1x1 + ··· + βkxk + ∼ N(x β, σ ) | {z } |{z} linear form x0β N(0,σ2) noise The model can be reformulate in terms of • distribution of the response: y | x ∼ N(µ, σ2), and • dependence of the mean on the predictors: µ = E(y | x) = x0β Dr. Guangliang Chen | Mathematics & Statistics, San Jos´e State University3/24 Generalized Linear Models (GLMs) beta=(1,2) 5 4 3 β0 + β1x b y 2 y 1 0 −1 0.0 0.2 0.4 0.6 0.8 1.0 x x Dr. Guangliang Chen | Mathematics & Statistics, San Jos´e State University4/24 Generalized Linear Models (GLMs) Generalized linear models (GLM) extend linear regression by allowing the response variable to have • a general distribution (with mean µ = E(y | x)) and • a mean that depends on the predictors through a link function g: That is, g(µ) = β0x or equivalently, µ = g−1(β0x) Dr. Guangliang Chen | Mathematics & Statistics, San Jos´e State University5/24 Generalized Linear Models (GLMs) In GLM, the response is typically assumed to have a distribution in the exponential family, which is a large class of probability distributions that have pdfs of the form f(x | θ) = a(x)b(θ) exp(c(θ) · T (x)), including • Normal - ordinary linear regression • Bernoulli - Logistic regression, modeling binary data • Binomial - Multinomial logistic regression, modeling general cate- gorical data • Poisson - Poisson regression, modeling count data • Exponential, Gamma - survival analysis Dr.
    [Show full text]
  • Logistic Regression, Dependencies, Non-Linear Data and Model Reduction
    COMP6237 – Logistic Regression, Dependencies, Non-linear Data and Model Reduction Markus Brede [email protected] Lecture slides available here: http://users.ecs.soton.ac.uk/mb8/stats/datamining.html (Thanks to Jason Noble and Cosma Shalizi whose lecture materials I used to prepare) COMP6237: Logistic Regression ● Outline: – Introduction – Basic ideas of logistic regression – Logistic regression using R – Some underlying maths and MLE – The multinomial case – How to deal with non-linear data ● Model reduction and AIC – How to deal with dependent data – Summary – Problems Introduction ● Previous lecture: Linear regression – tried to predict a continuous variable from variation in another continuous variable (E.g. basketball ability from height) ● Here: Logistic regression – Try to predict results of a binary (or categorical) outcome variable Y from a predictor variable X – This is a classification problem: classify X as belonging to one of two classes – Occurs quite often in science … e.g. medical trials (will a patient live or die dependent on medication?) Dependent variable Y Predictor Variables X The Oscars Example ● A fictional data set that looks at what it takes for a movie to win an Oscar ● Outcome variable: Oscar win, yes or no? ● Predictor variables: – Box office takings in millions of dollars – Budget in millions of dollars – Country of origin: US, UK, Europe, India, other – Critical reception (scores 0 … 100) – Length of film in minutes – This (fictitious) data set is available here: https://www.southampton.ac.uk/~mb1a10/stats/filmData.txt Predicting Oscar Success ● Let's start simple and look at only one of the predictor variables ● Do big box office takings make Oscar success more likely? ● Could use same techniques as below to look at budget size, film length, etc.
    [Show full text]
  • Bayesian Inference Chapter 4: Regression and Hierarchical Models
    Bayesian Inference Chapter 4: Regression and Hierarchical Models Conchi Aus´ınand Mike Wiper Department of Statistics Universidad Carlos III de Madrid Master in Business Administration and Quantitative Methods Master in Mathematical Engineering Conchi Aus´ınand Mike Wiper Regression and hierarchical models Masters Programmes 1 / 35 Objective AFM Smith Dennis Lindley We analyze the Bayesian approach to fitting normal and generalized linear models and introduce the Bayesian hierarchical modeling approach. Also, we study the modeling and forecasting of time series. Conchi Aus´ınand Mike Wiper Regression and hierarchical models Masters Programmes 2 / 35 Contents 1 Normal linear models 1.1. ANOVA model 1.2. Simple linear regression model 2 Generalized linear models 3 Hierarchical models 4 Dynamic models Conchi Aus´ınand Mike Wiper Regression and hierarchical models Masters Programmes 3 / 35 Normal linear models A normal linear model is of the following form: y = Xθ + ; 0 where y = (y1;:::; yn) is the observed data, X is a known n × k matrix, called 0 the design matrix, θ = (θ1; : : : ; θk ) is the parameter set and follows a multivariate normal distribution. Usually, it is assumed that: 1 ∼ N 0 ; I : k φ k A simple example of normal linear model is the simple linear regression model T 1 1 ::: 1 where X = and θ = (α; β)T . x1 x2 ::: xn Conchi Aus´ınand Mike Wiper Regression and hierarchical models Masters Programmes 4 / 35 Normal linear models Consider a normal linear model, y = Xθ + . A conjugate prior distribution is a normal-gamma distribution:
    [Show full text]
  • John Ashworth Nelder: 8 October 1924 – 7 August 2010
    John Ashworth Nelder: 8 October 1924 – 7 August 2010. John Nelder died on Saturday 7th August 2010 in Luton & Dunstable Hospital UK, where he was recovering from a fall. John was very active even at the age of 85, and retained the strong interest in our work – and statistics generally – that we will all remember with deep affection. However, he was becoming increasingly frail and it was a shock but perhaps, in retrospect, not a surprise to hear that he had died peacefully in his sleep. John was born on 8th October 1924 in Dulverton, Somerset, UK. He was educated at Blundell's School and at Sidney Sussex College, Cambridge where he read Mathematics (interrupted by war service in the RAF) from 1942-8, and then took the Diploma in Mathematical Statistics. Most of John’s formal career was spent as a statistician in the UK Agricultural Research Service. His first job, from October 1949, was at the newly set-up Vegetable Research Station, Wellesbourne UK (NVRS). Then, in 1968, he became Head of the Statistics Department at Rothamsted, and continued there until his first retirement in 1984. The role of statistician there was very conducive for John, not only because of his strong interests in biology (and especially ornithology), but also because it allowed him to display his outstanding skill of developing new statistical theory to solve real biological problems. At NVRS, John developed the theory of general balance to provide a unifying framework for the wide range of designs that are needed in agricultural research (see Nelder, 1965, Proceedings of the Royal Society, Series A).
    [Show full text]
  • Linear, Ridge Regression, and Principal Component Analysis
    Linear, Ridge Regression, and Principal Component Analysis Linear, Ridge Regression, and Principal Component Analysis Jia Li Department of Statistics The Pennsylvania State University Email: [email protected] http://www.stat.psu.edu/∼jiali Jia Li http://www.stat.psu.edu/∼jiali Linear, Ridge Regression, and Principal Component Analysis Introduction to Regression I Input vector: X = (X1, X2, ..., Xp). I Output Y is real-valued. I Predict Y from X by f (X ) so that the expected loss function E(L(Y , f (X ))) is minimized. I Square loss: L(Y , f (X )) = (Y − f (X ))2 . I The optimal predictor ∗ 2 f (X ) = argminf (X )E(Y − f (X )) = E(Y | X ) . I The function E(Y | X ) is the regression function. Jia Li http://www.stat.psu.edu/∼jiali Linear, Ridge Regression, and Principal Component Analysis Example The number of active physicians in a Standard Metropolitan Statistical Area (SMSA), denoted by Y , is expected to be related to total population (X1, measured in thousands), land area (X2, measured in square miles), and total personal income (X3, measured in millions of dollars). Data are collected for 141 SMSAs, as shown in the following table. i : 1 2 3 ... 139 140 141 X1 9387 7031 7017 ... 233 232 231 X2 1348 4069 3719 ... 1011 813 654 X3 72100 52737 54542 ... 1337 1589 1148 Y 25627 15389 13326 ... 264 371 140 Goal: Predict Y from X1, X2, and X3. Jia Li http://www.stat.psu.edu/∼jiali Linear, Ridge Regression, and Principal Component Analysis Linear Methods I The linear regression model p X f (X ) = β0 + Xj βj .
    [Show full text]
  • Strength in Numbers: the Rising of Academic Statistics Departments In
    Agresti · Meng Agresti Eds. Alan Agresti · Xiao-Li Meng Editors Strength in Numbers: The Rising of Academic Statistics DepartmentsStatistics in the U.S. Rising of Academic The in Numbers: Strength Statistics Departments in the U.S. Strength in Numbers: The Rising of Academic Statistics Departments in the U.S. Alan Agresti • Xiao-Li Meng Editors Strength in Numbers: The Rising of Academic Statistics Departments in the U.S. 123 Editors Alan Agresti Xiao-Li Meng Department of Statistics Department of Statistics University of Florida Harvard University Gainesville, FL Cambridge, MA USA USA ISBN 978-1-4614-3648-5 ISBN 978-1-4614-3649-2 (eBook) DOI 10.1007/978-1-4614-3649-2 Springer New York Heidelberg Dordrecht London Library of Congress Control Number: 2012942702 Ó Springer Science+Business Media New York 2013 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. Exempted from this legal reservation are brief excerpts in connection with reviews or scholarly analysis or material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Duplication of this publication or parts thereof is permitted only under the provisions of the Copyright Law of the Publisher’s location, in its current version, and permission for use must always be obtained from Springer.
    [Show full text]
  • Bayesian Methods: Review of Generalized Linear Models
    Bayesian Methods: Review of Generalized Linear Models RYAN BAKKER University of Georgia ICPSR Day 2 Bayesian Methods: GLM [1] Likelihood and Maximum Likelihood Principles Likelihood theory is an important part of Bayesian inference: it is how the data enter the model. • The basis is Fisher’s principle: what value of the unknown parameter is “most likely” to have • generated the observed data. Example: flip a coin 10 times, get 5 heads. MLE for p is 0.5. • This is easily the most common and well-understood general estimation process. • Bayesian Methods: GLM [2] Starting details: • – Y is a n k design or observation matrix, θ is a k 1 unknown coefficient vector to be esti- × × mated, we want p(θ Y) (joint sampling distribution or posterior) from p(Y θ) (joint probabil- | | ity function). – Define the likelihood function: n L(θ Y) = p(Y θ) | i| i=1 Y which is no longer on the probability metric. – Our goal is the maximum likelihood value of θ: θˆ : L(θˆ Y) L(θ Y) θ Θ | ≥ | ∀ ∈ where Θ is the class of admissable values for θ. Bayesian Methods: GLM [3] Likelihood and Maximum Likelihood Principles (cont.) Its actually easier to work with the natural log of the likelihood function: • `(θ Y) = log L(θ Y) | | We also find it useful to work with the score function, the first derivative of the log likelihood func- • tion with respect to the parameters of interest: ∂ `˙(θ Y) = `(θ Y) | ∂θ | Setting `˙(θ Y) equal to zero and solving gives the MLE: θˆ, the “most likely” value of θ from the • | parameter space Θ treating the observed data as given.
    [Show full text]
  • Generalized Linear Models
    Generalized Linear Models A generalized linear model (GLM) consists of three parts. i) The first part is a random variable giving the conditional distribution of a response Yi given the values of a set of covariates Xij. In the original work on GLM’sby Nelder and Wedderburn (1972) this random variable was a member of an exponential family, but later work has extended beyond this class of random variables. ii) The second part is a linear predictor, i = + 1Xi1 + 2Xi2 + + ··· kXik . iii) The third part is a smooth and invertible link function g(.) which transforms the expected value of the response variable, i = E(Yi) , and is equal to the linear predictor: g(i) = i = + 1Xi1 + 2Xi2 + + kXik. ··· As shown in Tables 15.1 and 15.2, both the general linear model that we have studied extensively and the logistic regression model from Chapter 14 are special cases of this model. One property of members of the exponential family of distributions is that the conditional variance of the response is a function of its mean, (), and possibly a dispersion parameter . The expressions for the variance functions for common members of the exponential family are shown in Table 15.2. Also, for each distribution there is a so-called canonical link function, which simplifies some of the GLM calculations, which is also shown in Table 15.2. Estimation and Testing for GLMs Parameter estimation in GLMs is conducted by the method of maximum likelihood. As with logistic regression models from the last chapter, the generalization of the residual sums of squares from the general linear model is the residual deviance, Dm 2(log Ls log Lm), where Lm is the maximized likelihood for the model of interest, and Ls is the maximized likelihood for a saturated model, which has one parameter per observation and fits the data as well as possible.
    [Show full text]
  • Robust Bayesian General Linear Models ⁎ W.D
    www.elsevier.com/locate/ynimg NeuroImage 36 (2007) 661–671 Robust Bayesian general linear models ⁎ W.D. Penny, J. Kilner, and F. Blankenburg Wellcome Department of Imaging Neuroscience, University College London, 12 Queen Square, London WC1N 3BG, UK Received 22 September 2006; revised 20 November 2006; accepted 25 January 2007 Available online 7 May 2007 We describe a Bayesian learning algorithm for Robust General Linear them from the data (Jung et al., 1999). This is, however, a non- Models (RGLMs). The noise is modeled as a Mixture of Gaussians automatic process and will typically require user intervention to rather than the usual single Gaussian. This allows different data points disambiguate the discovered components. In fMRI, autoregressive to be associated with different noise levels and effectively provides a (AR) modeling can be used to downweight the impact of periodic robust estimation of regression coefficients. A variational inference respiratory or cardiac noise sources (Penny et al., 2003). More framework is used to prevent overfitting and provides a model order recently, a number of approaches based on robust regression have selection criterion for noise model order. This allows the RGLM to default to the usual GLM when robustness is not required. The method been applied to imaging data (Wager et al., 2005; Diedrichsen and is compared to other robust regression methods and applied to Shadmehr, 2005). These approaches relax the assumption under- synthetic data and fMRI. lying ordinary regression that the errors be normally (Wager et al., © 2007 Elsevier Inc. All rights reserved. 2005) or identically (Diedrichsen and Shadmehr, 2005) distributed. In Wager et al.
    [Show full text]
  • Principal Component Analysis (PCA) As a Statistical Tool for Identifying Key Indicators of Nuclear Power Plant Cable Insulation
    Iowa State University Capstones, Theses and Graduate Theses and Dissertations Dissertations 2017 Principal component analysis (PCA) as a statistical tool for identifying key indicators of nuclear power plant cable insulation degradation Chamila Chandima De Silva Iowa State University Follow this and additional works at: https://lib.dr.iastate.edu/etd Part of the Materials Science and Engineering Commons, Mechanics of Materials Commons, and the Statistics and Probability Commons Recommended Citation De Silva, Chamila Chandima, "Principal component analysis (PCA) as a statistical tool for identifying key indicators of nuclear power plant cable insulation degradation" (2017). Graduate Theses and Dissertations. 16120. https://lib.dr.iastate.edu/etd/16120 This Thesis is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. Principal component analysis (PCA) as a statistical tool for identifying key indicators of nuclear power plant cable insulation degradation by Chamila C. De Silva A thesis submitted to the graduate faculty in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Major: Materials Science and Engineering Program of Study Committee: Nicola Bowler, Major Professor Richard LeSar Steve Martin The student author and the program of study committee are solely responsible for the content of this thesis. The Graduate College will ensure this thesis is globally accessible and will not permit alterations after a degree is conferred.
    [Show full text]
  • Generalized Linear Models Outline for Today
    Generalized linear models Outline for today • What is a generalized linear model • Linear predictors and link functions • Example: estimate a proportion • Analysis of deviance • Example: fit dose-response data using logistic regression • Example: fit count data using a log-linear model • Advantages and assumptions of glm • Quasi-likelihood models when there is excessive variance Review: what is a (general) linear model A model of the following form: Y = β0 + β1X1 + β2 X 2 + ...+ error • Y is the response variable • The X ‘s are the explanatory variables • The β ‘s are the parameters of the linear equation • The errors are normally distributed with equal variance at all values of the X variables. • Uses least squares to fit model to data, estimate parameters • lm in R The predicted Y, symbolized here by µ, is modeled as µ = β0 + β1X1 + β2 X 2 + ... What is a generalized linear model A model whose predicted values are of the form g(µ) = β0 + β1X1 + β2 X 2 + ... • The model still include a linear predictor (to right of “=”) • g(µ) is the “link function” • Wide diversity of link functions accommodated • Non-normal distributions of errors OK (specified by link function) • Unequal error variances OK (specified by link function) • Uses maximum likelihood to estimate parameters • Uses log-likelihood ratio tests to test parameters • glm in R The two most common link functions Log • used for count data η η = logµ The inverse function is µ = e Logistic or logit • used for binary data µ eη η = log µ = 1− µ The inverse function is 1+ eη In all cases log refers to natural logarithm (base e).
    [Show full text]