iForecast (version 1.0.4)

rollingWindows: Rolling timeframe for time series anaysis

Description

It extracts time stamp from a timeSeries object and separates the time into in-sample training and out-of-sample validation ranges.

Usage

rollingWindows(x,estimation="18m",by = "6m")

Arguments

x

The time series matrix of dependent variable, with timeSeries or zoo format.

estimation

The range of insample estimation period, the default is 18 months(18m), where the k-fold cross-section is performed.

by

The range of out-of-sample validation/testing period, the default is 6 months(6m).

Value

window

The time labels of from and to

.

Details

This function is similar to the backtesting framework in portfolio analysis. Rolling windows fixes the origin and the training sample grows over time, moving windows can be achieved by placing window() on dependent variable at each iteration.

Examples

# NOT RUN {
data(macrodata)
dep=macrodata[,1,drop=FALSE]
ind=macrodata[,-1,drop=FALSE]

timeframe=rollingWindows(dep,estimation="300m",by="6m")
FROM=timeframe$from
TO=timeframe$to

type=c("none","trend","season","both")[1]
models=c("rf","rpart","svm","knn","nnet")[4]
t0=1
# you may change this to "for (to in 1: length(TO))" to perform rolling CV
#output = ttsCaret(y=dep, x=ind, arOrder=c(1), xregOrder=c(1), method=models,
#                   tuneLength =10, train.end=as.character(TO[t0]), type=type)
#if(t0==length(TO)) {
#  testData=window(output$dataused,start=TO[t0],end=end(Y))[-1,]
#} else {testData=window(output$dataused,start=TO[t0],end=TO[t0+1])[-1,]}


#P1=iForecast(Model=output,newdata=testData,type="staticfit") # static forecasts generated
#P2=iForecast(Model=output,newdata=testData,type="recursive") # dynamic forecasts generated

#tail(cbind(testData[,1],P1),10)
#tail(cbind(testData[,1],P2),10)


# }