svd_CV: Model selection of a svd model using missing value based CV...

In YipengUva/RpESCA: R implementaion of pESCA models with various concave penalties

Description

This function implemented a missing value based CV model selection approach. First, ratio_mis percent elements are randomly selected as missing values. After that a EM-SVD model is constructed to estimate the prediction error. The details of this function can be found in https://arxiv.org/abs/1902.06241.

Usage

svd_CV(X, K = 3, Rs = 1:15, opts = list(), ratio_mis = 0.1)

Arguments

X	a quantitative data set
K	K-fold cross validation
Rs	the searching range of the number of components
opts	a list contains the setting for the algorithm. tol_obj: tolerance for relative change of hist_obj, default:1E-6; maxit: max number of iterations, default: 1000;
ratio_mis	the propotion of missing values

Value

This function returns a list contains

• Rs: the number of PCs used for model selection

• cvErrors: K-fold cross validation errors

Examples

## Not run: svd_CV(X,K=3,Rs = 1:15,opts=list(),ratio_mis=0.1)

YipengUva/RpESCA documentation built on July 2, 2019, 6:41 p.m.