Kickstart Your Probabilistic Forecasting
summary
Introduces two practical probabilistic forecasting methods — the Level Set Forecaster and Quantile Regression Forest — and explains how they provide uncertainty estimates without adding unnecessary complexity. Covers when to use each model as a straightforward way to kickstart probabilistic forecasts for time series or related prediction problems.
AI-written summary · about
chapters (11)
- 0:00Probabilistic forecasting intro
- 1:56Point forecasts limits
- 3:00Forecast fundamentals
- 4:27Evaluating forecasts
- 6:54Switch to probabilistic
- 7:56Practical models start
- 8:38Level set forecaster
- 13:51Quantile regression forest
- 18:09Comparing approaches
- 21:03Key takeaways
- 22:15Q&A
related talks
34mPrediction intervals
46mPanel about probabilistic forecasting
37mQuantifying Uncertainty in Time Series Forecasting
29mProbabilistic demand forecasting at scale
32mLeveraging Conformal Prediction for Calibrated Probabilistic Time Series Forecasts
30mKnowing what you don’t know matters- Uncertainty-aware model rating
27mModelling the extreme using Quantile Regression
29mTime Series Forecasting Techniques in Python