normalize.max: Max normalization

In gecko515/HEMDAG: Hierarchical Ensemble Methods for Directed Acyclic Graphs

Description

Function to normalize the scores of a flat scores matrix per class.

Usage

normalize.max(S)

Arguments

S	matrix with the raw non normalized scores. Rows are examples and columns are classes.

Details

The scores of each class are normalized by dividing the score values for the maximum score of that class. If the max score of a class is zero, no normalization is needed, otherwise NaN value will be printed as results of 0 out of 0 division.

Value

A score matrix with the same dimensions of S, but with scores max/normalized separately for each class.

Examples

data(scores);
maxnorm <- normalize.max(S);

gecko515/HEMDAG documentation built on Oct. 18, 2019, 6:34 a.m.