Bias in error rate estimates in discriminant analysis when stepwise variable selection is employed
Authors:
- 1 January 1991
- journal article
- research article
- Published by Taylor & Francis in Communications in Statistics - Simulation and Computation
- Vol. 20 (1) , 1-22
Abstract
Little attention has been given to properties of error rate estimates obtained from stepwise-selected discriminant analyses. We present simulation results that describe the effects of several factors on the error estimates' bias. Six estimators are compared; most can give substantially biased estimates when used in conjunction with stepwise discriminant analysis. Holdout methods are recommended because they are nearly unbiased and have desirable stability propertiesKeywords
This publication has 12 references indexed in Scilit:
- Cross-Validation, the Jackknife, and the Bootstrap: Excess Error Estimation in Forward Logistic RegressionJournal of the American Statistical Association, 1986
- Monte Carlo study of forward stepwise discrimination based on small samplesComputers & Mathematics with Applications, 1986
- Error-Rate Estimation in Discriminant AnalysisTechnometrics, 1985
- Multivariate Observations1984
- A Leisurely Look at the Bootstrap, the Jackknife, and Cross-ValidationThe American Statistician, 1983
- On the Relationship between the F Test and the Overall Error Rate for Variable Selection in Two-Group Discriminant Analysis1980
- Bias in Wilks' A in Stepwise Discriminant AnalysisTechnometrics, 1980
- Comparison of Stopping Rules in Forward Stepwise Discriminant AnalysisJournal of the American Statistical Association, 1979
- On Expected Probabilities of Misclassification in Discriminant Analysis, Necessary Sample Size, and a Relation with the Multiple Correlation Coefficient1968
- Estimation of Error Rates in Discriminant AnalysisTechnometrics, 1968