/IJF_AUSGDP

Reconciliation of quarterly Australian GDP forecasts from Income and Expenditure sides

Primary LanguageR

IJF_AUSGDP

Forecast combination based reconciliation of the quarterly Australian GDP from Income and Expenditure sides (Di Fonzo and Girolimetto, 2024)

The forecast reconciliation is performed using FoReco 0.2.2.

Files

Note: the analysis folders are distinct according to two different hierarchies. TYPE stands for either Income/INC or Expenditure/EXP.

  • BaseForecasts: Base forecasts as in Athanasopoulos et al. (2019) and naive forecasts (SA and SRW):
    • TYPE/TYPE_features_snaive.R: Seasonal Random Walk (SRW) and ARIMA base forecasts;
    • TYPE/TYPE_features_sa.R: Seasonal Average (SA) base forecasts.
  • Reconciliation:
    • TYPE/ARIMA:
      • TYPE_cslccd_bCCCred.R: LCC with ARIMA bts (exogenous)
      • TYPE_cslccd_bCCCred_endo.R: LCC with ARIMA bts (endogenous)
      • TYPE_cslccd_bLCC.R: LxCC with ARIMA bts (x = level number, exogenous)
      • INC_cslccd_bLCCendo.R: LxCC with ARIMA bts (x = level number, endogenous)
      • TYPE_cslccd_mean_red.R: LCC with SRW bts (exogenous)
      • TYPE_cslccd_mean_red_endo.R: LCC with SRW bts (endogenous)
      • TYPE_cslccd_mean.R: CCC with SRW bts
      • TYPE_cslccd_mLCC.R: LxCC with SRW bts (x = level number, exogenous)
      • TYPE_cslccd_mLCCendo.R: LxCC with SRW bts (x = level number, endogenous)
      • TYPE_cslccd_sa.R: CCC with SA bts
      • TYPE_cslccd.R: CCC with ARIMA bts and CCCH
      • TYPE_htsrec_mean.R: cross-sectional reconciliation with SRW bts
      • TYPE_htsrec.R: cross-sectional reconciliation with ARIMA bts
      • TYPE_CCCmix_red.R: sample average of LCC (SRW + ARIMA bts, exogenous)
      • TYPE_CCCmix_red_endo.R: sample average of LCC (SRW + ARIMA bts, endogenous)
      • TYPE_CCCmix.R: sample average of CCC (SRW + ARIMA bts, exogenous)
      • TYPE_CCCmix_endo.R: sample average of CCC (SRW + ARIMA bts, endogenous)
      • TYPE_csmix.R: sample average of the cross-sectional reconciliation (SRW + ARIMA bts)
    • TYPE/function: functions for TYPE/score/TYPE_scores_arima.R
    • TYPE/score/TYPE_scores_arima.R: returns the scores’ dataset

References

Athanasopoulos, G., Gamakumara, P., Panagiotelis, A., Hyndman, R.J., Affan, M., 2019. Hierarchical Forecasting, in: Fuleky, P. (Ed.), Macroeconomic Forecasting in the Era of Big Data. Springer, Cham, pp. 689–719. doi:10.1007/978-3-030-31150-6_21

Di Fonzo, T. and Girolimetto D. (2024). Forecast combination based forecast reconciliation: insights and extensions, International Journal of Forecasting, 40(2), pp. 490-514. doi:10.1016/j.ijforecast.2022.07.00