Calculate_CV_Error: Parameter tuning function #' For each candidate parameter...

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

Description

Parameter tuning function #' For each candidate parameter value, calculate the CV error resulting from growing a forest on TRAIN1, predicting TRAIN2, and downweighting contributions in TRAIN2 based on either RF residuals or RFL residuals using OOB predictions when TRAIN2 was predicted using RF grown on TRAIN2. In cross validation, this is done for each candidate value, within each fold, within each rep.

Usage

Calculate_CV_Error(OOBWeights1, PredWeights1, TRAINY, samp, parvec, ind,
  tol = 10^-4, BisqwtRF, BisqwtRFL, OutlierInd)

Arguments

OOBWeights1	OOB prediction weights for training cases in TRAIN1
PredWeights1	Prediction weights for predicting test (validation) cases using training cases
TRAINY	responses for all training cases
samp	list of indexes of test (or validation) cases (i.e. TRAIN2)
parvec	vector of candidate values for tuning parameter alpha
ind	index of parameter tuning vector to work on
tol	maximal change in interation for LOWESSRF weights in cross validation
BisqwtRF	Weight to be applied to each test or validation case when using cross validation to set tuning parameter, using RF outliers for downweighting
OutlierInd	Vector of zeros and ones indicating whether training cases came from contaminating distribution

Value

Returns a 6 by 3 matrix with errors. Rows indicate type of weighting applied to errors in TRAIN2. 1-MSE without downweighting outliers in CV error 2-MAPE without downweighting outliers in CV error 3-MSE downweighting outliers according to BisqwtRF 4-MAPE downweighting outliers according to BisqwtRF 5-MSE downweighting outliers according to BisqwtRFL 6-MAPE downweighting outliers according to BisqwtRFL Columns represent error for all cases (1), only outliers (2), only nonoutliers(3)

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