Automated univariate forecasting using the StatsForecast library

The StatsForecast library by Nixtla, introduced in early 2022, has gained quite a popularity since its initial release. Nixtla offers several specialized forecasting libraries, including MLForecast, which enables machine learning-based forecasting using standard machine learning models; NeuralForecast, which implements deep learning approaches for time series prediction; HierarchicalForecast, which handles multi-level time series structures with various reconciliation methods; and TimeGPT, which provides the proprietary GPT-based forecasting capabilities.The StatsForecast library offers a collection of popular univariate time series forecasting models, such as ARIMA, Theta, GARCH, ARCH, and exponential smoothing, with automatic variants. According to Nixtla’s website, their StatsForecast contains the “fastest and most accurate implementations of AutoArima, AutoETS, AutoCES, MSTL, and Theta in Python.”...