functime: a next generation ML forecasting library powered by Polars
summary
functime is a Python library that trains a single estimator across a panel to forecast thousands of time series on a laptop by using Polars (Rust + Rayon) for multithreaded, zero-copy feature engineering. It implements TS‑fresh–style feature extractors with Rust kernels, uses global forecasting (LightGBM/GBDT or linear with categorical entity features), offers backtesting, diagnostics and conformal prediction intervals, and reports 1–2 orders of magnitude speedups in feature engineering and CV so you avoid clusters.
AI-written summary · about
chapters (6)
- 0:00Introduction
- 0:13Forecasting challenges
- 11:25Live demo and features
- 22:45Diagnostics and evaluation
- 26:11Conclusion
- 27:02Q&A session
related talks
28mHow I used Polars to build built functime, a next gen ML forecasting library
45mUncertainty estimation at scale with functime
21mPolars: A highly optimized dataframe library
48mFKLearn, A functional machine learning library
42mRobust Automated Forecasting in Python and R
40mMulti-Series Forecasting at Scale with StatsForecast
51mFuture: Friendly Parallel Processing in R for Everyone
42mRobust Automated Forecasting in Python and R