ACEMS Forecasting Workshop

Date

26 July 2018

Venue

University of Melbourne

 

Date: 26 July 2018
Location: University of Melbourne

This page is for people enrolled in my ACEMS half-day workshop.

Prerequisites

Please bring your own laptop with a recent version of R and RStudio installed, along with the fpp2 package and its dependencies.

Participants will be assumed to be familiar with basic statistical tools such as multiple regression, but no knowledge of time series or forecasting will be assumed.

Reference

Online textbook on forecasting

Need help with R?

Program

1:00 - 1:45 Forecast evaluation Slides
1:45 - 2:30 ARIMA models Slides
2:30 - 3:00 Break
3:00 - 4:00 Dynamic regression Slides
4:00 - 5:00 Hierarchical forecasting Slides

 

Lab Session 1

  1. For the first four lab sessions, we will use the qcement data (Quarterly Australian Portand Cement production, 1956–2014). Plot the data using autoplot().
  2. Split the data into a training set and a test set of 4 years. We will apply models to the training set, and compare the forecasts on the test set. Use window() to split the data.
  3. Compute a seasonal naïve forecast applied to the training data, and plot the results. Use snaive() to produce the forecasts.
  4. Test if the residuals are white noise using the checkresiduals() function. What do you conclude?

Lab Session 2

  1. Compare the forecasts of the four benchmark methods to the test data using the accuracy() command.
  2. What do you conclude?

Lab Session 3

  1. For the qcement data, fit a suitable ARIMA model for the logged data using auto.arima(). You can use the argument lambda=0 in the auto.arima() function to take the logs. That way, the forecasts will be on the original scale.
  2. Does this model pass the residuals check?
  3. How does it compare to the benchmark models on the test data?

Lab Session 4

  1. Download quarterly GDP data from the Australian Macro Database. To make this a time series:

    ausgdp <- ts(read.csv("gdpcknaaoq.csv")[,1],
  start=c(1959,3), frequency=4)

  2. Fit a dynamic regression model to the logged qcement data with GDP as a predictor variable. Make sure you use the same time periods for both variables.

  3. How do the results compare with the ARIMA model fitted earlier?

Lab Session 5

  1. The visnights data set contains quarterly visitor nights for various regions of Australia. To turn this into an hts object:

    library(hts)
tourism.hts <- hts(visnights, characters = c(3, 5))

  2. Generate forecasts of the bottom level series using ARIMA models, and sum them for “bottom-up” forecasts:

    visnightsfc <- forecast(tourism.hts, method='bu', fmethod='arima', h=8)

autoplot(aggts(tourism.hts, level=0)) +
  autolayer(aggts(visnightsfc, level=0), lty=2)
autoplot(aggts(tourism.hts, level=1)) +
  autolayer(aggts(visnightsfc, level=1), lty=2)
autoplot(aggts(tourism.hts, level=2)) +
  autolayer(aggts(visnightsfc, level=2), lty=2)

    Do the forecasts look reasonable?

  3. Now use optimally reconciled forecasts:

    visnightsfc2 <- forecast(tourism.hts, fmethod='arima', h=8)

autoplot(aggts(tourism.hts, level=0)) +
  autolayer(aggts(visnightsfc2, level=0), lty=2)
autoplot(aggts(tourism.hts, level=1)) +
  autolayer(aggts(visnightsfc2, level=1), lty=2)
autoplot(aggts(tourism.hts, level=2)) +
  autolayer(aggts(visnightsfc2, level=2), lty=2)

autoplot(aggts(visnightsfc2, level=0)) +
  autolayer(aggts(visnightsfc, level=0), lty=2)
autoplot(aggts(visnightsfc2, level=1)) +
  autolayer(aggts(visnightsfc, level=1), lty=2)
autoplot(aggts(visnightsfc2, level=2)) +
  autolayer(aggts(visnightsfc, level=2), lty=2)

    What difference has the reconciliation step made?

Further resources for forecasting