Efficient approximate k-fold and leave-one-out cross-validation for ridge regression

摘要: In model building and evaluation, cross-validation is a frequently used resampling method. Unfortunately, this method can be quite time consuming. article, we discuss an approximation that much faster in generalized linear models Cox' proportional hazards with ridge penalty term. Our based on Taylor expansion around the estimate of full model. way, all cross-validated estimates are approximated without refitting The tuning parameter now chosen these approximations optimized less time. most accurate when approximating leave-one-out results for large data sets which originally computationally demanding situation. order to demonstrate method's performance, it will applied several microarray sets. An R package penalized, implements method, available CRAN.