0

Im looking for the best way to calculate the MAD, MAPE, MSE in R for a Holt-Winters and ARIMA forecast model. Below is my code for my models and any guidance would be awesome

Holt-Winters Model and Prediction:

#Holt-Winters Exponential Smoothing
forecast.mean <- HoltWinters(Abbeville$Incoming.Examinations,gamma=FALSE)


#Predicting ahead the next 12 months
forecast.predict <- predict(forecast.mean, n.ahead = 12, prediction.interval = TRUE)

ARIMA model and Forecast

ARIMA_forecast <- auto.arima(x=Abbeville$Incoming.Examinations)

ARIMA_nextforecasting <- forecast(ARIMA_forecast, h=12)
J. McCraiton
  • 159
  • 1
  • 2
  • 10
  • 2
    Use the `accuracy` function from the forecast package. – Rob Hyndman Nov 26 '17 at 00:37
  • Thank you, It worked for my ARIMA model, but not my Holts-Winter model. I receive this error: Error in accuracy.default(forecast.predict) : First argument should be a forecast object or a time series. This is what forecast.predict looks like: forecast.predict <- predict(forecast.mean, n.ahead = 12, prediction.interval = TRUE) – J. McCraiton Nov 26 '17 at 01:13
  • Use the forecast function to make predictions, not predict. – Rob Hyndman Nov 26 '17 at 01:55
  • Perfect, thank you that worked. – J. McCraiton Nov 26 '17 at 01:58

0 Answers0