sim

Simulated Australian tourism dataset (folder: Australia_data/sim)

Note: all the output *.csv files are inputs for lhf.Rmd file to display figures and tables. (Tables: 7 and 8 - Figures: 11, 12, 15 and 16)

Simulating the series:

Run seriessimulation.R file. It simulates all the simulated series we used with different forecast horizon, noise, level and number of series.

• input: TourismData_v3.csv
• output: actual.sim.noise.FH.csv, sim.304.csv, sim.608.csv, sim.1520.csv, sim.3040.csv

ETS and ARIMA:

Run ETS-ARIMA-noiseFH.R - changing the added noise (0.01, 0.1, 0.5, 1) and forecast horizon (12, 24, 36, 48) and ETS-ARIMA-levelsNS.Rfiles - changing number of levels (8, 10, 12, 18) and number of series (304, 608, 1520, 3040) (both rolling and fixed origin methods).

• input: TourismData_v3.csv
• output: fc.fix.sim.ets.arima.levelNS8.csv, fc.rolling.sim.ets.arima.levelNS8.csv, fc.fix.sim.ets.arima.levelNS10.csv, fc.rolling.sim.ets.arima.levelNS10.csv, fc.fix.sim.ets.arima.levelNS12.csv, fc.rolling.sim.ets.arima.levelNS12.csv, fc.fix.sim.ets.arima.levelNS18.csv, fc.rolling.sim.ets.arima.levelNS18.csv, fc.fix.noise0.01.FH.csv, fc.rolling.noise0.01.FH.csv, fc.fix.noise0.1.FH.csv, fc.rolling.noise0.1.FH.csv, fc.fix.noise0.5.FH.csv, fc.rolling.noise0.5.FH.csv,fc.fix.noise1.FH.csv and fc.rolling.noise1.FH.csv

Note: After running ETS-ARIMA-levelsNS.R file we need to run cleaning-results.R to clean the results anf get above outputs.

OLS:

Run OLS-fix-noiseFH.R, OLS-rolling-noiseFH.R, OLS-fix-levelsNS.R, OLS-rolling-levelsNS.R files. For rolling origin approach, the function is OLSmodel in OLSmodel_se.R file. For fixed origin approach, the function is olsfc in olsfc_se.R file. For computing reconciliation matrix we also need to run smatrix.R file. Then, to compute the OLS results, first we neet to run OLSmodel_se.R, olsfc_se.R and smatrix.R.

• input: TourismData_v3.csv
• output: OLS.fix.levelNS.csv, OLS.rolling.levelNS.csv, fc.OLS.fix.noise.FH.csv and fc.OLS.rolling.noise.FH.csv