INFLUENCE DIAGNOSTICS FOR THE CROSS-VALIDATED SMOOTHING PARAMETER IN SPLINE SMOOTHING

This article addresses the problem of influence in estimating the smoothing parameter when fitting a univariate smoothing spline. Diagnostics are presented that can identify observations that locally influence the choice of smoothing parameter by generalized cross-validation, either through their case weights or observed responses. Generalized cross-validation and the diagnostic methods are given a frequency interpretation and illustrated with examples.