trainDirectFit: Training with direct fit only

In sfeuerriegel/caret.ts: Time series models for "caret" package

Description

This function is an alternative training function with performs a direct fit only, i.e. it avoids any kind of cross-validation or any form of repeated executions. It is thus beneficial to use this training function when working with objects that solely require a direct fit, such as a least squares, ARMA or ARIMA. It thereby considerably reduces runtime compare to the default initialization of trainControl.

Usage

trainDirectFit(...)

Arguments

...	Optional parameters passed on to trainControl.

Value

An object that controls the training inside train.

Examples

library(forecast)

data(WWWusage) # from package "forecast"
df <- data.frame(y = as.numeric(WWWusage))

library(caret)

lm <- train(y ~ 1, data = df, method = "lm", trControl = trainDirectFit())
summary(lm)
RMSE(predict(lm, df), df)

arima <- train(y ~ 1, data = df, method = arima_model(1, 1, 1), trControl = trainDirectFit())
summary(arima)
RMSE(predict(arima, df), df)

sfeuerriegel/caret.ts documentation built on May 29, 2019, 8:01 p.m.