MakeLowessPred: Predict RFLOWESS Tuning

In AndrewjSage/RFLOWESS: A robust approach to random forest regression

Description

This is a function to be called inside TuneMultifoldCV. It makes the RFLOWESS predictions for a specific alpha and records them in a matrix LPREDERR.

Usage

MakeLowessPred(OOBWeights, PredWeights, TRAINY, TEST, tol = tol, ind, parvec,
  method = "Tukey")

Arguments

OOBWeights	matrix of training cases with response in last column
PredWeights	Vector of zeros and ones indicating whether training cases came from contaminating distribution
TRAINY	number of folds to perform in cross validation
TEST	Test data
tol	maximal change in interation for LOWESSRF weights in cross validation
ind	index of parameter vector to use in tuning
parvec	vector of candidate values for tuning parameter alpha
method	should Tukey or Huber weighting function be used?
ndsize	nodesize random forest tuning parameter for cross validation

Value

Returns: LPREDERR: A length 2 vector containing MSPE and MAPE on test data using alpha parameter specified LWeights: A ntest by ntrain matrix containing LOWESSRF weights for test cases LIter: Number of iterations until convergence of LOWESSRF algorithm

AndrewjSage/RFLOWESS documentation built on May 26, 2019, 6:38 a.m.