proto: Extract the prototype from each variable group

In OHPL: Ordered Homogeneity Pursuit Lasso for Group Variable Selection

Extract the prototype from each variable group

Description

Extracts the prototype from each variable group.

Usage

proto(X, y, groups, type = c("max", "median"), mu = NULL)

Arguments

X	Predictor matrix.
y	Response matrix with one column.
groups	An group index vector containing the group number each variable belongs to. For example: c(1, 1, 1, 1, 1, 2, 2, 2, ...). Variable groups can be generated by the Fisher optimal partition algorithm implemented in FOP().
type	The rule for extracting the prototype. Possible options are "max" and "median".
mu	The mean value of y for standardization. Default is NULL, which uses the sample mean of y.

Value

The prototypes (variable index) extracted from each group (cluster).

OHPL documentation built on Sept. 11, 2024, 7:05 p.m.