RFSampleScheme: RFSampleScheme: A Random Forest function.

In jstarling1/starlib: All of my useful functions.

Description

This function builds a vector of sample sizes to be used as the sampsize parameter in the random forest function. The vector contains sample sizes for each class level.

Usage

RFSampleScheme(.totalObs, .minObsPerClass, .nKnown, .propClassKnown, .nUnknown,
  .propClassUnknown)

Arguments

totalObs	The number of observatiosn desired for the entire data set, all classes together. Defaults to NULL.
minObsPerClass	The minimum number of observations allowed per class. Defaults to NULL.
nKnown	Number of observataions in the data set with known class labels. Defaults to NULL.
nUnknown	Number of observataions in the data set with unknown class labels. Defaults to NULL.
propClassUnknown	Estimated class proportiosnin the data set with unknown class labels. Defaults to NULL.

Value

A vector containing sampsize class proportions to use in the RandomForest model.

Author(s)

Robert Saw

jstarling1/starlib documentation built on May 20, 2019, 2:12 a.m.