My notes on the evaluation of forecasting methods (multi-part series).

Part 1: Measures of forecasting accuracy

Measures of forecasting accuracy

Work in progress…

1. Model selection and evaluation
2. Diagnostic tools
3. Statistical tests of forecast performance
4. Choosing the appropriate level of aggregation when evaluating forecasts
5. Dealing with constrained historical observed demand

References

General

📃 Review articles:

• Hewamalage, et al. (2022) - Forecast Evaluation for Data Scientists: Common Pitfalls and Best Practices
  • 🗃️ arXiv version→ https://arxiv.org/abs/2203.10716
• Petropoulos et al. (2022) - Forecasting: theory and practice
  • 🗃️ arXiv version → https://doi.org/10.48550/arXiv.2012.03854

📚 Books and book chapters:

• Manu Joseph (2022) - Modern Time Series Forecasting with Python
  • Chapters of interest:
    • 18 Evaluating Forecasts – Forecast Metrics
    • 19 Evaluating Forecasts – Validation Strategies
• Rob J. Hyndman (2021) - Forecasting: Principles and Practice
  • Chapters of interest:
    • 5.8 Evaluating point forecast accuracy
    • 5.10 Time series cross-validation

Measures of forecasting accuracy

• Sai, C., Davydenko, A., & Shcherbakov, M. (2018) - List of publications on measures of forecasting accuracy