Seuck Heun Song

An extensive comparison of recent classification tools applied to microarray data

Since most classi%cation articles have applied a single technique to a single gene expression dataset, it is crucial to assess the performance of each method through a comprehensive comparative study. We evaluate by extensive comparison study extending Dudoit et al. (J. Amer. Statist. Assoc. 97 (2002) 77) the performance of recently developed classi%cation methods in microarray experiment, and provide the guidelines for %nding the most appropriate classi%cation tools in various situations. We extend their comparison in three directions: more classi%cation methods (21 methods), more datasets (7 datasets) and more gene selection techniques (3 methods). Our comparison study shows several interesting facts and provides the biologists and the biostatisticians some insights into the classi%cation tools in microarray data analysis. This study also shows that the more sophisticated classi%ers give better performances than classical methods such as kNN, DLDA, DQDA and the choice of gene selection method has much e>ect on the performance of the classi%cation methods, and thus the classi%cation methods should be considered together with the gene selection criteria. c 2004 Elsevier B.V. All rights reserved.

The unbalanced nested error component regression model

Abstract This paper considers a nested error component model with unbalanced data and proposes simple analysis of variance (ANOVA), maximum likelihood (MLE) and minimum norm quadratic unbiased estimators (MINQUE)-type estimators of the variance components. These are natural extensions from the biometrics, statistics and econometrics literature. The performance of these estimators is investigated by means of Monte Carlo experiments. While the MLE and MINQUE methods perform the best in estimating the variance components and the standard errors of the regression coefficients, the simple ANOVA methods perform just as well in estimating the regression coefficients. These estimation methods are also used to investigate the productivity of public capital in private production.

Testing Panel Data Regression Models with Spatial Error Correlation

This paper derives several Lagrange Multiplier tests for the panel data regression model with spatial error correlation. These tests draw upon two strands of earlier work. The first is the LM tests for the spatial error correlation model discussed in Anselin (1988, 1999) and Anselin, Bera, Florax and Yoon (1996), and the second is the LM tests for the error component panel data model discussed in Breusch and Pagan (1980) and Baltagi, Chang and Li (1992). The idea is to allow for both spatial error correlation as well as random region effects in the panel data regression model and to test for their joint significance. Additionally, this paper derives conditional LMtests, which test for random regional effects given the presence of spatial error correlation. Also, spatial error correlation given the presence of random regional effects. These conditional LM tests are an alternative to the one directional LM tests that test for random regional effects ignoring the presence of spatial error correlation or the one directional LM tests for spatial error correlation ignoring the presence of random regional effects. We argue that these joint and conditional LM tests guard against possible misspecification. Extensive Monte Carlo experiments are conducted to study the performance of these LM tests as well as the corresponding Likelihood Ratio tests.

Testing for heteroskedasticity and spatial correlation in a random effects panel data model

A panel data regression model with heteroskedastic as well as spatially correlated disturbances is considered, and a joint LM test for homoskedasticity and no spatial correlation is derived. In addition, a conditional LM test for no spatial correlation given heteroskedasticity, as well as a conditional LM test for homoskedasticity given spatial correlation, are also derived. These LM tests are compared with marginal LM tests that ignore heteroskedasticity in testing for spatial correlation, or spatial correlation in testing for homoskedasticity. Monte Carlo results show that these LM tests, as well as their LR counterparts, perform well, even for small N and T. However, misleading inferences can occur when using marginal, rather than joint or conditional LM tests when spatial correlation or heteroskedasticity is present.

Multiclass sparse logistic regression for classification of multiple cancer types using gene expression data

Monitoring gene expression profiles is a novel approach to cancer diagnosis. Several studies have showed that the sparse logistic regression is a useful classification method for gene expression data. Not only does it give a sparse solution with high accuracy, it provides the user with explicit probabilities of classification apart from the class information. However, its optimal extension to more than two classes is not obvious. In this paper, we propose a multiclass extension of sparse logistic regression. Analysis of five publicly available gene expression data sets shows that the proposed method outperforms the standard multinomial logistic model in prediction accuracy as well as gene selectivity.

SIMPLE LM TESTS FOR THE UNBALANCED NESTED ERROR COMPONENT REGRESSION MODEL

ABSTRACT This paper derives several Lagrange Multiplier tests for the unbalanced nested error component model. Economic data with a natural nested grouping include firms grouped by industry; or students grouped by schools. The LM tests derived include the joint test for both effects as well as the test for one effect conditional on the presence of the other. The paper also derives the standardized versions of these tests, their asymptotic locally mean most powerful version as well as their robust to local misspecification version. Monte Carlo experiments are conducted to study the performance of these LM tests.

Testing for overdispersion in a censored Poisson regression model

In this article, we investigate the efficiency of score tests for testing a censored Poisson regression model against censored negative binomial regression alternatives. Based on the results of a simulation study, score tests using the normal approximation, underestimate the nominal significance level. To remedy this problem, bootstrap methods are proposed. We find that bootstrap methods keep the significance level close to the nominal one and have greater power uniformly than does the normal approximation for testing the hypothesis.

BLUP in the nested panel regression model with serially correlated errors

Abstract The best linear unbiased predictor for the panel data regression model with serially correlated nested error components is derived. Furthermore, performance of the predictor is compared with the other predictors using the study of productivity of public capital in private production based on a panel of 28 states over the period 1970–1986. The estimators whose predictions are compared include OLS, nested effects ML estimator ignoring serial correlation and nested effects ML estimator accounting for the serial correlation. Based on prediction mean square error (PMSE) forecast performance, it is crucial to take into account nested effects as well as serial correlation.

LM Tests for the Unbalanced Nested Panel Data Regression Model with Serially Correlated Errors

This paper derives several Lagrange Multiplier tests for the unbalanced nested error component model with serially correlated remainder disturbances. The problems of overtesting and undertesting for serial correlation and zero random group and nested subgroup effects are considered. The joint test extends the earlier work of Breush and Pagan [1980] and King and Wu [1997] to the unbalanced nested error component regression model with serially correlated errors. Additionally, conditional LM tests, asymptotically local mean most powerful (LMMP) tests; modified Rao-Score tests that guard against local misspecification are proposed for this model. These generalize the work of Baltagi and Li [1995], Rahman and King [1998] and Bera and Yoon [1993]. Finally, Monte Carlo experiments are conducted to study the performance of these LM tests.

The efficiency of the sample mean in a linear regression model when errors follow a first-order moving average process

Abstract In this paper a linear regression model is considered with the sample mean to be estimated. Errors are assumed to follow a MA(1) process. We derive the efficiency function for the OLS estimator relative to the best linear unbiased estimator (Aitken) for finite sample size. In the case of positive correlation among the errors a greatest lower bound for the efficiency is established.