A high correlation between predicted and criterion values in cross-validation may not always be a sufficient indicator of goodness of prediction. An alternative indicator of goodness of prediction is reviewed in terms of what kind of prediction is sought.
Get full access to this article
View all access options for this article.
References
1.
Motoyama, T.Comparison of three methods for prediction: The least square method, ridge regression, and equal weighting. Doctoral dissertation, Iowa State University, 1978.
2.
Nunnally, J.C.Psychometric theory (2nd ed.). New York: McGraw-Hill, 1978.
