Variable selection in discriminant analysis: measuring the influence of individual cases

Steel, S. J.; Louw, Nelmarie

doi:10.1016/s0167-9473(01)00012-3

Cited by 6 publications

(3 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If we recalculate (2) using the data D(,) and only the variables in V(D(,)), we obtain RV(Dw)(t). Steel and Louw (2001) propose as a measure of the influence of data case k, taking selection into account. Note that Q,(t) quantifies the relative change in the accuracy of the post-selection posterior probability estimates if case k is omitted from the training data.…”

Section: The Q and Hb Methods For Dealing With Selection Influential mentioning

confidence: 99%

See 1 more Smart Citation

Variable selection in discriminant analysis in the presence of outliers

Stell

Louw

2001

Proceedings of the 23rd International Conference on Information Technology Interfaces, 2001. ITI 2001.

Self Cite

View full text Add to dashboard Cite

A n important problem in discriminant analysis that has received little attention in the literature, is the effect of outliers when variable selection is the first step in the anabsis.In this paper we consider two methodi of dealing with this problem. n e classification performance of these methods, and the classrfication performance If possible outliers are ignored, are compared in a simulation study. In the configrations srudied, weJind that none of the approaches consistently ourperforms the others.

show abstract

Section: The Q and Hb Methods For Dealing With Selection Influential mentioning

confidence: 99%

“…The third approach uses the technique proposed by Steel and Louw (2001), in which identification of outliers and selection of variables are integrated. This is referred to as method Q below.…”

Section: Introductionmentioning

confidence: 99%

Variable selection in discriminant analysis in the presence of outliers

Stell

Louw

2001

Proceedings of the 23rd International Conference on Information Technology Interfaces, 2001. ITI 2001.

Self Cite

View full text Add to dashboard Cite

show abstract

“…OtherSteel and Louw (2001) present formulas for exact calculation of bootstrap estimates of expected prediction error for k-nearest-neighbor classification, and propose a "weighted k-NN classifier" based on resampling ideas Velilla and Barrio (1994). propose a data transformation for nonnormal distributions in a linear or quadratic PDA context.Definition Ordinal scale:Device for taking weights in the Vatican.EXERCISES1.…”

mentioning

confidence: 99%

Answers to Exercises

2005

Wiley Series in Probability and Statistics

View full text Add to dashboard Cite

List of Figures1.1 Classification of multivariate methods. 2.1 Distance in a plane. 5.1 LDF plot of group centroids for Baumann study. 5.2 Plot of group centroids in LDF space. 6.1 Plot of Wilks values versus best subset size for the 3-group Ethington data. 7.1 LDF plot of group centroids. 7.2 MANOVA and descriptive discriminant analysis.8.1 LDF plot for the three school levels. 9.1 Two dimensional plot of adjusted group centroids. 12.1 Distance in a plane. 12.2 Graphical representations of two density functions. 17.1 Total group L-O-O hit rate versus best-subset size for the 3-group Ethington data. 20.1 Predictive discriminant analysis. xix M. S. Poudevigne (Clayton State University).There are several students who attended The University of Georgia and gave us considerable assistance in preparing this second edition. Jason Lemley was very patient and persistent in helping to process some of the text. Veena Dhankher and In-Suk Kim were very helpful in providing computer assistance for several data sets used in this book. And special thanks to the students in the 2005 Fall section of the multivariate class, particularly Hye-Jeong Choi and Chen-Yao Koa, who persistently requested further clarification of the text and identified omissions and mistypes in an earlier draft of this book.We would also like to thank our University of Georgia Colleague, Seock-Ho Kim, for his assistance in the use of the LaTex text formating software.The patience and efforts of Wiley editors, including Christine Punzo, Susanne Steitz, and Stephen Quigley, are greatly appreciated.xxvii Rank of score on X 19.2.1

show abstract