convergence_plot: convergence_plot

In horserule: Flexible Non-Linear Regression with the HorseRule Algorithm

Description

Can be used to check model convergence.

Usage

convergence_plot(model, Xtest, ytest, burnin = 0)

Arguments

model	list containing a model of class "hs_rulefit".
Xtest	Out of bag sample to check error.
ytest	response of test data.
burnin	Number of samples disregarded as burnin.

Details

Convergence is checked by the convergence of the prediction error on unseen test data, to find a suitable number of iterations, in the spirit of gradient boosting. To check convergence on the Training data just use training X and y instead of Xtest and ytest.

Examples

library(MASS)
data(Boston)
#Split in train and test data
N = nrow(Boston)
train = sample(1:N, 400)
Xtrain = Boston[train,-14]
ytrain = Boston[train, 14]
Xtest = Boston[-train, -14]
ytest = Boston[-train, 14]

hrres = HorseRuleFit(X = Xtrain, y=ytrain,
                    thin=1, niter=100, burnin=10,
                    L=5, S=6, ensemble = "both", mix=0.3, ntree=100,
                    intercept=FALSE, linterms=1:13, ytransform = "log",
                    alpha=1, beta=2, linp = 1, restricted = 0)

#Check the model convergence out of sample
convergence_plot(hrres, Xtest, ytest, burnin = 10)

Example output

horserule documentation built on May 2, 2019, 10:04 a.m.