avg.arit: arithmetic average of two models

In fizlaz/DUN: Team DUN Kaggle milestone

Description

This version is made for 5-label classification. Inputs are prediction probability matricesof two classifiers.

Usage

avg.arit(mat1, mat2, label = train$popularity, iter = 101)

Arguments

mat1	Matrix of first classifier class probabiliets.
mat2	Matrix of second classifier class probabiliets.
label	True class labels. Default train set popularity column.
iter	Number of weight combinations, default 101 (0.01 increment)

Value

An ordered matrix with weight combinations and matching accuracies

fizlaz/DUN documentation built on May 16, 2019, 1:13 p.m.