Time-series forecasting

You have a series of observations (temperature, sales, stock returns, traffic) over time, and you use machine learning to predict/forecast how the value will change in the future.

Notes

Useful resources

Videos

• Time Series Forecasting with XGBoost - Use python and machine learning to predict energy consumption
  • https://www.youtube.com/watch?v=vV12dGe_Fho
• Time Series Forecasting with XGBoost - Advanced Methods
  • https://www.youtube.com/watch?v=z3ZnOW-S550

Books

• Forecasting: Principles and Practice (3rd edition)
  • https://otexts.com/fpp3/
• Forecasting: Principles and Practice (2nd edition)
  • https://otexts.com/fpp2/
• A Very Short Course on Time Series Analysis
  • https://bookdown.org/rdpeng/timeseriesbook/

Tutorials

• https://www.tensorflow.org/tutorials/structured_data/time_series

Links

• https://en.wikipedia.org/wiki/Makridakis_Competitions
• https://robjhyndman.com/hyndsight/forecasting-competitions/
• https://www.statsmodels.org/stable/tsa.html
• https://clickhouse.com/blog/forecasting-using-clickhouse
• https://github.com/blue-yonder/tsfresh

Leave a Comment