predict_ensemble: Generate class prediction probabilities

In ryanholbrook/bayeserror: Estimating Bayes Error in Classification Tasks

Description

This function will generate class prediction probabilities on a dataset from a list of classifiers using a specified resampling method.

Usage

predict_ensemble(x, y, classifiers, v = 10L)

Arguments

x	A matrix with observations as rows and features as columns.
y	A vector of 0/1 response classes
classifiers	A list of classifier functions. Each function should be of the form f(x, y, newdata) and have the attribute shortname to identify its predictions.
v	For k-fold cross-validation, use an integer to specify the number of folds. For instance, v = 10 means to use 10-fold CV. For a holdout split, use a number between 0.0 to 1.0 specifying the percent to be held out to make predictions. For in-sample predictions, use v = 0.

Value

A three-dimensional array containing class prediction probabilities. Its dimesions are as (observation, class, classifier). The last entry on the third dimension is "AVE", the predictions for the ensemble classifier.

ryanholbrook/bayeserror documentation built on Jan. 12, 2020, 6:25 p.m.