09-26, 13:25–14:00 (Europe/Amsterdam), Apollo
Probabilistic forecasting is essential, but choosing the right method is tricky. This talk introduces two lesser-known models — Level Set Forecaster and Quantile Regression Forest — that help you kickstart probabilistic forecasting without unnecessary complexity.
You want a probabilistic forecast, but where do you begin? There’s no single best model, and the landscape of probabilistic methods can feel overwhelming. From Bayesian models to neural networks, the variety is huge, and the learning curve steep.
This talk helps you skip the theory maze and get hands-on with two lesser-known, practical models that are easy to implement with Python:
- Level Set Forecaster: a simple wrapper that turns any existing point forecast into a probabilistic forecast using a nonparametric approach. Ideal if you already have a point model in place.
- Quantile Regression Forest: a powerful variation of Random Forests that outputs a probabilistic forecast, offering sample-specific uncertainty estimates.
Although powerful, these models are often buried in academic papers, making them hard to approach. This talk cuts through the complexity and offers a practical, implementation-focused guide you can apply right away.
This talk is ideal for data scientists who are comfortable with point prediction models and want to start applying probabilistic forecasting.