All Tools
M
DataFreeOpen Source
MLFORECAST
Scalable machine learning forecasts with lag features and sklearn models
Apache-2.0
ABOUT
Gradient boosting and sklearn models forecast well once you engineer lags, rolling stats, and calendar features, but that plumbing does not scale across thousands of series. MLForecast builds those features, fits local or distributed estimators, and produces panel forecasts so teams reuse LightGBM or XGBoost instead of hand-rolled lag pipelines.
INSTALL
pip install mlforecastINTEGRATION GUIDE
1. Forecast many SKUs with LightGBM or XGBoost on lag and date features
2. Scale panel forecasting to Spark or Dask clusters without rewriting feature code
3. Combine exogenous regressors with recursive or direct multi-step strategies
4. Benchmark ML models against StatsForecast and NeuralForecast in the Nixtla stack
TAGS
pythontime-seriesforecastingmachine-learningsklearnnixtlaspark