aGrum/pyAgrum in 2021

As the end of the year approaches, it's time for us to look back on our hard year of work. 💪 First of all, some numbers 📊: in 2021, aGrUM/pyAgrum has been downloaded more than 160K times on all platforms (anaconda, pip, others), that is to say more than 10K downloads per month on average! According to google scholar, it has also been cited 21 times in research papers.

In all, we have released 15 versions including 4 minor ones. These releases allowed us to improve influence diagrams and credal networks, to propose a new module (skbn) for Bayes classifier compatible with sklearn and to add new functionalities around explainability such as Shapley values.

This year, we had to say goodbye to python 2.7 and versions <3.7 😢 but we have added support for arm, aarch64 and python 3.10 🤝. Finally, let's also note that pyAgrum is now compiled in c++17 !

Despite all these advances, the best is still to come and we give you an appointment next year for the arrival of version 1.0.0 (finally!) and for the aGrUM day. We will communicate on these topics very soon so stay tuned to learn more. Until then, have a great holiday season! 🎉