R functions for improved time-series forecasting
forecastingTimeSeriesCrossValidation.R defines functions to implement time-series cross-validation (aka rolling forecast windows) for a given ARIMA specification and ETS specification. These ideas were outlined by Rob Hyndman (Monash University), creater of the forecast package in R. Since creation of this script, Rob Hyndman has added similar functionality to the forecast package, described here. However, the script defined here includes more granular plotting output of performance across rolling windows.
forecastingPlotErrorsHistogram.R defines a function to take a vector of forecast errors and plot a histogram of errors including a blue line representing a standard normal distribution for comparison.