cv.plot: Crossfold Validation Performance Plot.

In gboris/blkbox: Data Exploration with Multiple Machine Learning Algorithms

Description

Compares the performance of each algorithm in a boxplot OR barplot. Each holdout will contribute at least one data point to each algorithm.

Usage

cv.plot(obj, metric = "AUROC", y_ranges = c(0, 1), title = "",
  type = "boxplot")

Arguments

obj	An object produced by the blkboxCV function.
metric	Which metric you wish to plot. Area under the Receiver operating curve = "AUROC", Accuracy = "ACC", Error rate = "ERR", Matthews correlation coefficient = "MCC", F-1 score = "F-1". default = c("AUROC")
y_ranges	is the y axis limits for the plot, defaults to c(0,1). Must be a numeric vector with two entries. Invalid for barplots.
title	the title to be adhered to the plot. Default is no title.
type	The plot can be either a barplot or boxplot. For the barplot the consensus performance is used, for a boxplot consensus is false. If only one performance measure is found for each algorithm then it will be forced to a barplot. default = "boxplot", unless data is unsupported.

Author(s)

Zachary Davies, Boris Guennewig

Examples

cv.plot(blkboxCV(...), metric = "AUROC", title = "Example Performance Barplot")

gboris/blkbox documentation built on Dec. 27, 2021, noon