I'm thrilled to announce the latest release of MetaTS, a #Python package that simplifies and accelerates global #timeseries #forecasting using #meta-learning.

You can find MetaTS on GitHub here: https://github.com/DrSasanBarak/metats

Meta-learning has emerged as a winning solution for recent time series #forecasting competitions, and MetaTS is designed to make #meta-learning more accessible to researchers and data scientists. With MetaTS, you can easily generate meta-features using automated feature extraction and deep unsupervised learning, implement base-forecaster models, and optimize meta-parameters using a flexible and customizable pipeline.

In addition to providing a user-friendly toolkit for meta-learning, MetaTS also unifies the available Python libraries that can be useful for time series forecasting. You can leverage the power of #Sktime, #Nixtla, #Darts, and other libraries to create base forecasters and explore different meta-model architectures, including #stacking and #ensembling.

I'm proud of what we've achieved with MetaTS, and I believe it can be a valuable resource for anyone looking to improve their time series forecasting #performance. The latest version of the package is available on GitHub, and we welcome any feedback or contributions to help make MetaTS even better.

This can not be done without a great dedication and contribution of my colleague @AmirabbasAsadi .

Be tuned about this project on my LinkedIn

Thank you for your support!