cowbell (version 0.1.0)

generateCowbell: Performs the segmented linear regression analysis generating the cowbell function.

Description

This function takes the cowbell definition that was created with generateCowbellConcept and performs a regression analysis. Additionally it also moves the breaking point to the maximal values of the independent variables to later on test for the significance of the breaking point. This function needs relatively long to compute as it uses a gradient based optimizer for optimizing a non - linear model.

Usage

generateCowbell(concept, table, iterations = 1000, learningRate = 0.01)

Arguments

concept

The previously in function generateCowbellConcept specified concept.

table

The table that at least contains the data for the dependent and the two independent variables specified in concept.

iterations

The number of iteration that should be done with the gradient optimizer.

learningRate

The step size that should be applied in the gradient optimizer.

Value

A list with the data, the model with and without breakpoint and the F-Statistics.

See Also

generateCowbellConcept

Examples

# Run a simplified anaylsis with 10 iterations only (to save time.)
concept<-generateCowbellConcept(Fun ~ Fluency * Absorption, 1, 9, 1, 7, 1, 7)
data(allFun)
test<-generateCowbell(concept, allFun, 10)