train.ts: Train function to support time series objects

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

Description

Extends the train function from the "caret" package to support time series objects.

Usage

## S3 method for class 'ts'
train(x, ...)

Arguments

x	Object of class ts.
...	Further arguments passed on to train for training.

Value

Trained model.

Examples

library(caret)

library(forecast)
data(WWWusage)

class(WWWusage)
str(WWWusage)

arima <- train(WWWusage, method = auto_arima_model(), trControl = trainDirectFit())
summary(arima)
arimaorder(arima$finalModel) # order of best model
RMSE(predict(arima, WWWusage), WWWusage) # in-sample RMSE

library(vars)
data(Canada)

class(Canada)
str(Canada)

arima <- train(x = Canada[, -2], y = Canada[, 2], 
               method = auto_arima_model(), trControl = trainDirectFit())
summary(arima)
arimaorder(arima$finalModel) # order of best model
RMSE(predict(arima, Canada[, -2]), Canada[, 2]) # in-sample RMSE

arima <- train(form = prod ~ ., data = Canada, 
               method = auto_arima_model(), trControl = trainDirectFit())
summary(arima)
arimaorder(arima$finalModel) # order of best model
RMSE(predict(arima, Canada), Canada[, "prod"]) # in-sample RMSE

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