Feature selection is a necessary critical stage in data mining process. There is always an arm race to build frameworks and libraries that ease and automate this process. In this chapter, an EvoloPy-FS framework is proposed, which is a Python open-source optimization framework that includes several well-regarded swarm intelligence (SI) algorithms. It is geared toward feature selection optimization problems. It is an easy to use, reusable, and adaptable framework. The objective of developing EvoloPy-FS is providing a feature selection engine to help researchers even those with less knowledge in SI in solving their problems and visualizing rapid results with a less programming effort. That is why the orientation of this work was to build an open-source, white-box framework, where algorithms and data structures are being explicit, transparent, and publicly available. EvoloPy-FS comes to continue our path for building an integrated optimization environment, which was started by the original EvoloPy for global optimization problems, then EvoloPy-NN for training multilayer perception neural network, and finally the new EvoloPy-FS for feature selection optimization. EvoloPy-FS is freely hosted on (www.evo-ml.com) with a helpful documentation.

Published Paper

Useful Links

• The source code can be found on GitHub here
• Published Paper can be found here