Abstract
The jackknife, multicrossvalidation, and analytical formulae methods are discussed and compared with respect to the estimation of shrinkage of the multiple coefficient of correlation. The validation and comparison of the techniques are based on estimates generated from a random data set and a prestructured data set. Multicrossvalidation is demonstrated to be a preferable technique to the jackknife and analytical formulae methods.

This publication has 4 references indexed in Scilit: