knn_weighted: Model weights for "weight_type == "knn"

In kmaurer/binsemble:

Description

Calculate the model weights using k-nearest neighbor accuracy weights when "weight_type" == "knn"

Usage

knn_weighted(train_data, test_data, M, scale = TRUE, knn_size = 10)

Arguments

train_data	Training data with predicted class columns from each model 1,...,M
scale	TRUE/FALSE for rescaling data (default: scale=TRUE)
n	The number of instances in the test data
neighbors	number of neighbors used in accuracy estimates for each model

Value

matris of weights nrow=number of test points, rcol=M=number of models

kmaurer/binsemble documentation built on May 7, 2019, 9:50 p.m.