Empirical Performance of Cross-Validation With Oracle Methods in a Genomics Context
Empirical Performance of Cross-Validation With Oracle Methods in a Genomics Context
When employing model selection methods with oracle properties such as the smoothly clipped absolute deviation (SCAD) and the Adaptive Lasso, it is typical to estimate the smoothing parameter by m-fold cross-validation, for example, m = 10. In problems where the true regression function is sparse and the signals large, such …