Highlights from the 1.7.5 release of StatsForecast.⚡ • Addition of the MFLES model • Wrapper for scikit-learn models to leverage exogenous features ✨ Thanks to @Tyler Blume for the addition of the MFLES model! The model has excellent performance, is very fast, supports exogenous features, and multiple seasonalities. The MFLES model is based on Gradient Boosted Time Series Decomposition, which treats traditional decomposition as the base estimator in the boosting process. It derives its name from some of the underlying estimators that can enter into the boosting procedure: a simple Median, Fourier terms, a simple/piecewise Linear trend, and Exponential Smoothing. 🎁 StatsForecast now supports scikit-learn models through the statsforecast.models.SklearnModel wrapper. This can help you leverage feature engineering and train one model per series, which can sometimes be better than training a single global model (as in mlforecast). See more in the guide linked in the comments. See more in the release notes and Sklearn guide and thanks to @José Morales for all his work on this release.

Congrats to @Tyler Blume 🎉 and kudos to @José Morales and Nixtla team. mfles is a great model and a nice addition to statsforecast’ growing stack.

Super excited about the release! Many thanks to @José Morales for dealing with me throughout the process!