DOCSLIB.ORG

Maximum Likelihood for Variance Estimation in High-Dimensional Linear Models

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Maximum Likelihood for Variance Estimation in High-Dimensional Linear Models Maximum Likelihood for Variance Estimation in High-Dimensional Linear Models Lee H. Dicker Murat A. Erdogdu Rutgers University Stanford University Abstract 1 INTRODUCTION Variance parameters | including the residual vari- We study maximum likelihood estimators ance, the proportion of explained variation, and, under (MLEs) for the residual variance, the signal- some interpretations (e.g. Dicker, 2014; Hartley & Rao, to-noise ratio, and other variance parame- 1967; Janson et al., 2015), the signal-to-noise ratio | ters in high-dimensional linear models. These are important parameters in many statistical models. parameters are essential in many statisti- While they are often not the primary focus of a given cal applications involving regression diagnos- data analysis or prediction routine, variance parame- tics, inference, tuning parameter selection for ters are fundamental for a variety of tasks, including: high-dimensional regression, and other appli- cations, including genetics. The estimators (i) Regression diagnostics, e.g. risk estimation (Bay- that we study are not new, and have been ati et al., 2013; Mallows, 1973). widely used for variance component estima- (ii) Inference, e.g. hypothesis testing and construct- tion in linear random-effects models. How- ing confidence intervals (Fan et al., 2012; Javan- ever, our analysis is new and it implies that mard & Montanari, 2014; Reid et al., 2013). the MLEs, which were devised for random- (iii) Optimally tuning other methods, e.g. tuning pa- effects models, may also perform very well rameter selection for high-dimensional regression in high-dimensional linear models with fixed- (Sun & Zhang, 2012). effects, which are more commonly studied (iv) Other applications, e.g. genetics (De los Campos in some areas of high-dimensional statistics. et al., 2015; Listgarten et al., 2012; Loh et al., The MLEs are shown to be consistent and 2015; Yang et al., 2010). asymptotically normal in fixed-effects mod- els with random design, in asymptotic set- This paper focuses on variance estimation in high- tings where the number of predictors (p) is dimensional linear models. We study maximum proportional to the number of observations likelihood-based estimators for the residual variance (n). Moreover, the estimators' asymptotic and other related variance parameters. The results in variance can be given explicitly in terms mo- this paper show that | under appropriate conditions ments of the Marˇcenko-Pastur distribution. | the maximum likelihood estimators (MLEs) out- A variety of analytical and empirical results perform several other previously proposed variance es- show that the MLEs outperform other, pre- timators for high-dimensional linear models, in theory viously proposed estimators for variance pa- and in numerical examples, with both real and simu- rameters in high-dimensional linear models lated data. with fixed-effects. More broadly, the results in this paper illustrate a strategy for draw- The MLEs studied here are not new. Indeed, they ing connections between fixed- and random- are widely used for variance components estimation in effects models in high dimensions, which may linear random-effects models (e.g. Searle et al., 1992). be useful in other applications. However, the analysis in this paper is primarily con- cerned with the performance of the MLEs in fixed- effects models, which are more commonly studied in Appearing in Proceedings of the 19th International Con- some areas of modern high-dimensional statistics (e.g. ference on Artificial Intelligence and Statistics (AISTATS) B¨uhlmann& Van de Geer, 2011). Thus, from one per- 2016, Cadiz, Spain. JMLR: W&CP volume 41. Copyright spective, this paper may be viewed as an article on 2016 by the authors. model misspecification: We show that a classical es- Running heading title breaks the line timator for random-effects models may also be used where the design matrix X was highly structured, i.e. effectively in fixed-effects models. More broadly, the xij ∼ N(0; 1) were all iid (a detailed description of the results in this paper illustrate a strategy for drawing numerical simulations used to generate Figure 1 may connections between fixed- and random-effects mod- be found in the Supplementary Material). In general, els in high dimensions, which may be useful in other structured X methods work best when the entries of X applications; for instance, a similar strategy has been are uncorrelated. Indeed, by generating X with highly employed in Dicker's (2016) analysis of ridge regres- correlated columns and taking β to be very sparse, it sion. is easy to generate plots similar to those in Figure 1, where the structured X methods are badly biased and 1.1 Related Work the structured β methods perform well. On the other hand, if the design matrix X is correlated, it may be There is an abundant literature on variance compo- possible to improve the performance of structured X nents estimation for random-effects models, going back methods by \decorrelating" X (i.e. right-multiplying to at least the 1940s (Crump, 1946; Satterthwaite, X by a suitable positive definite matrix); this is not 1946). More recently, a substantial literature has de- pursued in detail here, but may be an interesting area veloped on the use of random-effects models in high for future research. throughput genetics (e.g. De los Campos et al., 2015; Listgarten et al., 2012; Loh et al., 2015; Yang et al., The method labeled \MLE" in Figure 1 is the main 2010). Maximum likelihood estimators for variance focus of this paper. The other methods depicted in parameters, like those studied in this paper, are widely Figure 1 that rely on structured X are \MM," \Eigen- Prism," and \AMP." MM is a method-of-moments es- used in this research area. A variety of questions about 2 the use of MLEs for variance parameters in modern timator for σ0 proposed in (Dicker, 2014); EigenPrism genetics remain unanswered; including fundamental was proposed in (Janson et al., 2015) and is the so- lution to a convex optimization problem; AMP is a questions about whether fixed- or random-effects mod- 2 els are more appropriate (De los Campos et al., 2015). method for estimating σ0 based on approximate mes- Our work here focuses primarily on fixed-effects mod- sage passing algorithms and was proposed in (Bay- els, and thus differs from much of this recent work on ati et al., 2013). In Figure 1, each of the structured genetics. However, Theorem 2 (see, in particular, the X methods is evidently unbiased. Additionally, it is discussion in Section 4.2) and the data analysis in Sec- clear that the variability of MLE is uniformly smaller tion 6 below suggest that fixed-effects analyses could than that of MM and EigenPrism (detailed results are possibly offer improved power over random-effects ap- reported in Table S1 of the Supplementary Material). proaches to variance estimation, in some settings. The AMP method is unique among the four structured X methods, because it is the only one that can adapt Switching focus to high-dimensional linear models to sparsity in β; consequently, while the variability of with fixed-effects, previous work on variance estima- MLE is smaller than that of AMP when β is not sparse tion has typically been built on one of two sets of (e.g. when sparsity is 40%), the AMP estimator has assumptions: Either (i) the underlying regression pa- smaller variance when β is sparse. This is certainly rameter [β in the model (1) below] is highly structured an attractive property of the AMP estimator. How- (e.g. sparse) or (ii) the design matrix X = (xij) is ever, little is known about the asymptotic distribution highly structured [e.g. xij are iid N(0; 1)]. Notable of the AMP estimator. By contrast, one important work on variance estimation under the \structured β" feature of the MLE is that it is asymptotically nor- assumption includes (Chatterjee & Jafarov, 2015; Fan mal in high dimensions and its asymptotic variance is et al., 2012; Sun & Zhang, 2012). Bayati et al. (2013), given by the simple formula in Theorem 2. This is 2 Dicker (2014), and Janson et al. (2015) have stud- a useful property for performing inference on σ0 and ied variance estimation in the \structured X" setting. related parameters, and for better understanding the This paper also focuses on the structured X setting. asymptotic performance of the MLE. As is characteristic of the structured X setting, many of the results in this paper require strong assumptions 1.2 Overview of the Paper on X; however, none of our results require any sparsity assumptions on β. Section 2 covers preliminaries. We introduce the sta- tistical model and define a bivariate MLE for the resid- Figure 1 illustrates some potential advantages of the ual variance and the signal-to-noise ratio | this is the structured X approach. In particular, it shows that main estimator of interest. We also give some addi- structured β methods for estimating the residual vari- tional background on the \structured X" model that ance σ2 can be biased, if β is not extremely sparse, 0 is studied here. while the structured X methods are effective regard- less of sparsity. Figure 1 was generated from datasets In Section 3, we present a coupling argument, which Lee H. Dicker, Murat A. Erdogdu 10%−sparse β 40%−sparse β 90%−sparse β 99.8%−sparse β MLE MLE MLE MLE MM MM MM MM EigenPrism EigenPrism EigenPrism EigenPrism AMP AMP AMP AMP Methods Methods Methods Methods Scaled−Lasso Scaled−Lasso Scaled−Lasso Scaled−Lasso RCV−Lasso RCV−Lasso RCV−Lasso RCV−Lasso 0.0 2.5 5.0 0.0 2.5 5.0 0.0 2.5 5.0 0.0 2.5 5.0 2 2 2 2 σ^ σ^ σ^ σ^ 2 Figure 1: Estimates of the residual variance σ0 from 500 independent datasets.
Load more

Recommended publications
  • 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]
  • Measures of Explained Variation and the Base-Rate Problem for Logistic Regression
    American Journal of Biostatistics 2 (1): 11-19, 2011 ISSN 1948-9889 © 2011 Science Publications Measures of Explained Variation and the Base-Rate Problem for Logistic Regression 1Dinesh Sharma, 2Dan McGee and 3B.M. Golam Kibria 1Department of Mathematics and Statistics, James Madison University, Harrisonburg, VA 22801 2Department of Statistics, Florida State University, Tallahassee, FL 32306 3Department of Mathematics and Statistics, Florida International University, Miami, FL 33199 Abstract: Problem statement: Logistic regression, perhaps the most frequently used regression model after the General Linear Model (GLM), is extensively used in the field of medical science to analyze prognostic factors in studies of dichotomous outcomes. Unlike the GLM, many different proposals have been made to measure the explained variation in logistic regression analysis. One of the limitations of these measures is their dependency on the incidence of the event of interest in the population. This has clear disadvantage, especially when one seeks to compare the predictive ability of a set of prognostic factors in two subgroups of a population. Approach: The purpose of this article is to study the base-rate sensitivity of several R2 measures that have been proposed for use in logistic regression. We compared the base-rate sensitivity of thirteen R2 type parametric and nonparametric statistics. Since a theoretical comparison was not possible, a simulation study was conducted for this purpose. We used results from an existing dataset to simulate populations with different base-rates. Logistic models are generated using the covariate values from the dataset. Results: We found nonparametric R2 measures to be less sensitive to the base-rate as compared to their parametric counterpart.
    [Show full text]
  • Principal Component Analysis and Optimization: a Tutorial Robert Reris Virginia Commonwealth University, [email protected]
    CORE Metadata, citation and similar papers at core.ac.uk Provided by VCU Scholars Compass Virginia Commonwealth University VCU Scholars Compass Statistical Sciences and Operations Research Dept. of Statistical Sciences and Operations Publications Research 2015 Principal Component Analysis and Optimization: A Tutorial Robert Reris Virginia Commonwealth University, [email protected] J. Paul Brooks Virginia Commonwealth University, [email protected] Follow this and additional works at: http://scholarscompass.vcu.edu/ssor_pubs Part of the Statistics and Probability Commons Creative Commons Attribution 3.0 Unported (CC BY 3.0) Recommended Citation Principal component analysis (PCA) is one of the most widely used multivariate techniques in statistics. It is commonly used to reduce the dimensionality of data in order to examine its underlying structure and the covariance/correlation structure of a set of variables. While singular value decomposition provides a simple means for identification of the principal components (PCs) for classical PCA, solutions achieved in this manner may not possess certain desirable properties including robustness, smoothness, and sparsity. In this paper, we present several optimization problems related to PCA by considering various geometric perspectives. New techniques for PCA can be developed by altering the optimization problems to which principal component loadings are the optimal solutions. This Conference Proceeding is brought to you for free and open access by the Dept. of Statistical Sciences and Operations Research at VCU Scholars Compass. It has been accepted for inclusion in Statistical Sciences and Operations Research Publications by an authorized administrator of VCU Scholars Compass. For more information, please contact [email protected]. http://dx.doi.org/10.1287/ics.2015.0016 Creative Commons License Computing Society 14th INFORMS Computing Society Conference This work is licensed under a Richmond, Virginia, January 11{13, 2015 Creative Commons Attribution 3.0 License pp.
    [Show full text]
  • Prediction of Stress Increase in Unbonded Tendons Using Sparse Principal Component Analysis
    Utah State University DigitalCommons@USU All Graduate Plan B and other Reports Graduate Studies 8-2017 Prediction of Stress Increase in Unbonded Tendons using Sparse Principal Component Analysis Eric Mckinney Utah State University Follow this and additional works at: https://digitalcommons.usu.edu/gradreports Part of the Applied Mathematics Commons, Applied Statistics Commons, and the Statistical Theory Commons Recommended Citation Mckinney, Eric, "Prediction of Stress Increase in Unbonded Tendons using Sparse Principal Component Analysis" (2017). All Graduate Plan B and other Reports. 1034. https://digitalcommons.usu.edu/gradreports/1034 This Creative Project is brought to you for free and open access by the Graduate Studies at DigitalCommons@USU. It has been accepted for inclusion in All Graduate Plan B and other Reports by an authorized administrator of DigitalCommons@USU. For more information, please contact [email protected]. PREDICTION OF STRESS INCREASE IN UNBONDED TENDONS USING SPARSE PRINCIPAL COMPONENT ANALYSIS by Eric McKinney A project submitted in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE in Statistics Approved: Yan Sun Marc Maguire Major Professor Committee Member Adele Cutler Daniel Coster Committee Member Committee Member UTAH STATE UNIVERSITY Logan, Utah 2017 ii Copyright c Eric McKinney 2017 All Rights Reserved iii ABSTRACT Prediction of Stress Increase in Unbonded Tendons using Sparse Principal Component Analysis by Eric McKinney, Master of Science Utah State University, 2017 Major Professor: Dr. Yan Sun Department: Mathematics and Statistics While internal and external unbonded tendons are widely utilized in concrete struc- tures, the analytic solution for the increase in unbonded tendon stress, ∆fps, is chal- lenging due to the lack of bond between strand and concrete.
    [Show full text]
  • Simple Linear Regression 80 60 Rating 40 20
    Simple Linear 12 Regression Material from Devore’s book (Ed 8), and Cengagebrain.com Simple Linear Regression 80 60 Rating 40 20 0 5 10 15 Sugar 2 Simple Linear Regression 80 60 Rating 40 20 0 5 10 15 Sugar 3 Simple Linear Regression 80 60 Rating 40 xx 20 0 5 10 15 Sugar 4 The Simple Linear Regression Model The simplest deterministic mathematical relationship between two variables x and y is a linear relationship: y = β0 + β1x. The objective of this section is to develop an equivalent linear probabilistic model. If the two (random) variables are probabilistically related, then for a fixed value of x, there is uncertainty in the value of the second variable. So we assume Y = β0 + β1x + ε, where ε is a random variable. 2 variables are related linearly “on average” if for fixed x the actual value of Y differs from its expected value by a random amount (i.e. there is random error). 5 A Linear Probabilistic Model Definition The Simple Linear Regression Model 2 There are parameters β0, β1, and σ , such that for any fixed value of the independent variable x, the dependent variable is a random variable related to x through the model equation Y = β0 + β1x + ε The quantity ε in the model equation is the “error” -- a random variable, assumed to be symmetrically distributed with 2 2 E(ε) = 0 and V(ε) = σ ε = σ (no assumption made about the distribution of ε, yet) 6 A Linear Probabilistic Model X: the independent, predictor, or explanatory variable (usually known).
    [Show full text]
  • Correlation and Regression Overview
    Introduction to Statistical Methods for Measuring “Omics” and Field Data Correlation and regression Overview • Correlation • Simple Linear Regression Correlation General Overview of Correlational Analysis • The purpose is to measure the strength of a linear relationship between 2 variables. • A correlation coefficient does not ensure “causation” (i.e. a change in X causes a change in Y) • X is typically the input, measured, or independent variable. • Y is typically the output, predicted, or dependent variable. • If X increases and there is a predictable shift in the values of Y, a correlation exists. General Properties of Correlation Coefficients • Values can range between +1 and -1 • The value of the correlation coefficient represents the scatter of points on a scatterplot • You should be able to look at a scatterplot and estimate what the correlation would be • You should be able to look at a correlation coefficient and visualize the scatterplot Interpretation • Depends on what the purpose of the study is… but here is a “general guideline”... • Value = magnitude of the relationship • Sign = direction of the relationship Correlation graph Strong relationships Weak relationships Y Y Positive correlation X X Y Y Negative correlation X X The Pearson Correlation Coefficient Correlation Coefficient • The correlation coefficient is a measure of the strength and the direction of a linear relationship between two variables. The symbol r represents the sample correlation coefficient. The formula for r is n∑xy − ∑x ∑ y r = ( )( ) . 2 2 n∑x 2 − (∑x) n∑ y 2 − (∑ y) • The range of the correlation coefficient is -1 to 1. If x and y have a strong positive linear correlation, r is close to 1.
    [Show full text]
  • Linear Regression
    Linear Regression Fall 2001 Professor Paul Glasserman B6014: Managerial Statistics 403 Uris Hall General Ideas of Linear Regression 1. Regression analysis is a technique for using data to identify relationships among vari- ables and use these relationships to make predictions. We will be studying linear re- gression, in which we assume that the outcome we are predicting depends linearly on the information used to make the prediction. Linear dependence means constant rate of increase of one variable with respect to another (as opposed to, e.g., diminishing returns). 2. Some motivating examples. (a) Suppose we have data on sales of houses in some area. For each house, we have com- plete information about its size, the number of bedrooms, bathrooms, total rooms, the size of the lot, the corresponding property tax, etc., and also the price at which the house was eventually sold. Can we use this data to predict the selling price of a house currently on the market? The first step is to postulate a model of how the various features of a house determine its selling price. A linear model would have the following form: selling price = β0 + β1 (sq.ft.) + β2 (no. bedrooms) +β3 (no. bath) + β4 (no. acres) +β5 (taxes) + error In this expression, β1 represents the increase in selling price for each additional square foot of area: it is the marginal cost of additional area. Similarly, β2 and β3 are the marginal costs of additional bedrooms and bathrooms, and so on. The intercept β0 could in theory be thought of as the price of a house for which all the variables specified are zero; of course, no such house could exist, but including β0 gives us more flexibility in picking a model.
    [Show full text]
  • Measures of Explained Variation in Gamma Regression Models
    COMMUN. STATIST.—SIMULA., 31(1), 61–73 (2002) REGRESSION AND TIME SERIES MEASURES OF EXPLAINED VARIATION IN GAMMA REGRESSION MODELS Martina Mittlbo¨ ck1,* and Harald Heinzl1,2,y 1Department of Medical Computer Sciences, University of Vienna, Spitalgasse 23, A-1090 Vienna, Austria 2Division of Biostatistics, German Cancer Research Center, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany ABSTRACT The common R2 measure provides a useful means to quantify the degree to which variation in the dependent variable can be explained by the covariates in a linear regression model. Recently, there have been various attempts to apply the defi- nition of the R2 measure to generalized linear models. This paper studies two different R2 measure definitions for the gamma regression model. These measures are related to deviance and sum-of-squares residuals. Depending on the sample size and the number of covariates fitted, so-called unadjusted R2 measures may be substantially inflated, and the use of adjusted R2 measures is then preferred. We study several known adjustments previously proposed for R2 mea- sures in regression models and illustrate the effect on the two unadjusted R2 measures for the gamma regression model. *E-mail: [email protected] yE-mail: [email protected] 61 Copyright & 2002 by Marcel Dekker, Inc. www.dekker.com FIRST PROOFS i:/Mdi/Sac/31(1)/Sac 31(1)-005.3d Statistics, Simulation and Computation (SAC) Paper: Sac 31(1)-005 62 MITTLBO¨ CK AND HEINZL Comparing the resulting measures with underlying population values, we find the best adjustment via simulation. Key Words: Adjusted R2 measures; Deviance; Sum-of- squares; Shrinkage; Degrees of freedom; Predictive power; Explained variation 1.
    [Show full text]
  • Linear Regression Using Ordinary Least Squares
    Simple Linear Regression Using Ordinary Least Squares Purpose: To approximate a linear relationship with a line. Reason: We want to be able to predict Y using X. Definition: The Least Squares Regression (LSR) line is the line with the smallest sum of square residuals smaller than any other line. That is, we want to measure closeness of the line to the points. The LSR line uses vertical distance from points to a line. A residual is the vertical distance from a point to a line. Equations: The true line for the population parameter: y = α + βx Note: We are trying to estimate this equation. We obtain a sample and estimate an approximation: yˆ = a + bx + E Estimate y Predict α = y-intercept Predict β = slope x = The independent variable. E = Epsilon or the error. These are random errors of measurement. Check the Assumptions: 1. No outliers. 2. Residuals follow a normal distribution with a mean = 0. 3. Residuals should be randomly scattered. Note: For #2, produce a histogram of the standardized residuals to see if they are normal. Hypotheses for the Correlation Coefficient (R): A measure of how close residuals are to the regression line. H0: ρ = 0 H1: ρ ≠ or < or > 0 Coefficient of Determination = R2 Range 0 – 1 0 ≤ r 2 ≤ 1 R2 = An effect size measure and yields the percentage of the variation in the Y values explained by X. Adjusted R2: Adjusts for R2’s upward bias and is a variance accounted for effect size measure. 2 Adj. R = .40 40% of variation in Y is explained by the regression line and dependent on X.
    [Show full text]
  • Statistical Analysis in JASP
    2nd Edition JASP v0.9.1 October 2018 Copyright © 2018 by Mark A Goss-Sampson. All rights reserved. This book or any portion thereof may not be reproduced or used in any manner whatsoever without the express written permission of the author except for the purposes of research, education or private study. CONTENTS PREFACE .................................................................................................................................................. 1 USING THE JASP INTERFACE .................................................................................................................... 2 DESCRIPTIVE STATISTICS ......................................................................................................................... 9 EXPLORING DATA INTEGRITY ................................................................................................................ 16 DATA TRANSFORMATION ..................................................................................................................... 23 ONE SAMPLE T-TEST ............................................................................................................................. 27 BINOMIAL TEST ..................................................................................................................................... 30 MULTINOMIAL TEST .............................................................................................................................. 33 CHI-SQUARE ‘GOODNESS-OF-FIT’ TEST............................................................................................
    [Show full text]
  • Choosing Sample Size
    Choosing Sample Size The purpose of this section is to describe three related methods for choosing sample size before data are collected -- the classical power method, the sample variation method and the population variation method. The classical power method applies to almost any statistical test. After presenting general principles, the discussion zooms in on the important special case of factorial analysis of variance with no covariates. The sample variation method and the population variation methods are limited to multiple linear regression, including the analysis of variance and covariance. Throughout, it will be assumed that the person designing the study is a scientist who will only be allowed to discuss results if a null hypothesis is rejected at some conventional significance level such as α = 0.05 or α = 0.01. Thus, it is vitally important that the study be designed so that scientifically interesting effects are likely to be be detected as statistically significant. The classical power method. The term "null hypothesis" has mostly been avoided until now, but it's much easier to talk about the classical power method if we're allowed to use it. Most statistical tests are based on comparing a full model to a reduced model. Under the reduced model, the values of population parameters are constrained in some way. For example, in a one-way ANOVA comparing σ2 three treatments, the parameters are µ1, µ2, µ3 and . The reduced model says that µ1=µ2=µ3. This is a constraint on the parameter values. The null hypothesis (symbolized H0) is a statement of how the parameters are constrained under the reduced model.
    [Show full text]
  • Principal Component Analysis and Optimization: a Tutorial Robert Reris Virginia Commonwealth University, [email protected]
    Virginia Commonwealth University VCU Scholars Compass Statistical Sciences and Operations Research Dept. of Statistical Sciences and Operations Publications Research 2015 Principal Component Analysis and Optimization: A Tutorial Robert Reris Virginia Commonwealth University, [email protected] J. Paul Brooks Virginia Commonwealth University, [email protected] Follow this and additional works at: http://scholarscompass.vcu.edu/ssor_pubs Part of the Statistics and Probability Commons Creative Commons Attribution 3.0 Unported (CC BY 3.0) Recommended Citation Principal component analysis (PCA) is one of the most widely used multivariate techniques in statistics. It is commonly used to reduce the dimensionality of data in order to examine its underlying structure and the covariance/correlation structure of a set of variables. While singular value decomposition provides a simple means for identification of the principal components (PCs) for classical PCA, solutions achieved in this manner may not possess certain desirable properties including robustness, smoothness, and sparsity. In this paper, we present several optimization problems related to PCA by considering various geometric perspectives. New techniques for PCA can be developed by altering the optimization problems to which principal component loadings are the optimal solutions. This Conference Proceeding is brought to you for free and open access by the Dept. of Statistical Sciences and Operations Research at VCU Scholars Compass. It has been accepted for inclusion in Statistical Sciences and Operations Research Publications by an authorized administrator of VCU Scholars Compass. For more information, please contact [email protected]. http://dx.doi.org/10.1287/ics.2015.0016 Creative Commons License Computing Society 14th INFORMS Computing Society Conference This work is licensed under a Richmond, Virginia, January 11{13, 2015 Creative Commons Attribution 3.0 License pp.
    [Show full text]