SampleSizeTest: Test the affect of changing sample size on model performance

In Tomhigg/fracCover: Generate and Validate Fractional Coverage products using Landsat/MODIS

Description

Test how changing a model sample size affect model performance. Function requires separate training and test data. Currently only parallel processing using doParalell is supported, will add a non parallel version soon

Usage

SampleSizeTest(train, trainingY, testY, testX, sizes, NoCores)

Arguments

train	A dataframe/matrix containing a dependent variable and set of independent (predictor) variable used for model training
trainingY	Name of the dependent variable
testY	A vector of dependent variable to be used for model validation (testing)
testX	A data frame of predictor variables used for testing, row numbers MUST corespond to the values in testY
sizes	A vector of samples sizes to be tested
NoCores	Number of Cores to used to parallel processing

Value

A dataframe of model performance statistics for each sample size test. Metrics returned include coefficient of determination (Rsq), root mean squared error (RMSE), mean error (ME), mean absolute error (MAE).

Tomhigg/fracCover documentation built on May 9, 2019, 5:11 p.m.