The bootstrap, jackknife and linear discriminant functions. (English) Zbl 0911.62050
Summary: Two linear discriminant functions are introduced based on a bias reduction technique, the jackknife technique, and a resampling technique, the bootstrap technique. These linear functions are compared with the linear discriminant function under normal and non-normal distributions using Monte Carlo methods. It is observed that the bootstrap discriminant function is superior to the others.
MSC:
62H30 | Classification and discrimination; cluster analysis (statistical aspects) |