Time Series Forecasting¶

AutoGluon can forecast the future values of multiple time series given the historical data and other related covariates. A single call to AutoGluon TimeSeriesPredictor’s fit() method trains multiple models to generate accurate probabilistic forecasts, and does not require you to manually deal with cumbersome issues like model selection and hyperparameter tuning.

Under the hood, AutoGluon combines various state of the art forecasting algorithms. These include established statical methods like ETS and ARIMA from StatsForecast, efficient tree-based forecasters like LightGBM based on AutoGluon-Tabular, flexible deep learning models like DeepAR and Temporal Fusion Transformer from GluonTS, and a pretrained zero-shot forecasting model, Chronos.

Check out the Quick Start Tutorial to learn how to make accurate forecasts in just 3 lines of code using AutoGluon.

Quick Start

Quick start tutorial on fitting models with time series datasets.

forecasting-quick-start.html

In-depth Tutorial

Detailed discussion of the time series forecasting capabilities in AutoGluon.

forecasting-indepth.html

Forecasting with Chronos

Zero-shot forecasting with pretrained Chronos time series models in AutoGluon.

forecasting-chronos.html

Model Zoo

List of available forecasting models in AutoGluon-TimeSeries.