Infosel is a feature selection library written in C++ that was designed to support the development of new algorithms based on information theory. Currently, Infosel++ contains about 60 different algorithms. Infosel++ can be obtained here An advantage of Infosel++ is the ability to use it from the Java and Matlab environment using this extension infosel.zip

For more details see:

• M. Blachnik, A. Bukowiec, M. Kordos, J. Biesiada Information Theory vs Correlation Based Feature Ranking Methods in Application to Metallurgical Problem Solving. LNCS 6113 pp. 289-298. 2010
• A. Kachel, J. Biesiada, M. Blachnik, W. Duch Infosel++: Information Based Feature Selection C++ Library. LNCS 6113 pp. 388-396. 2010
• M. Blachnik Comparison of Various Feature Selection Methods in Application to Prototype Best Rules. Advances in Intelligent and Soft Computing 57 pp. 257-264. Springer Verlag. 2009
• M. Blachnik, W. Duch, A. Kachel, J. Biesiada Feature Selection for High-Dimensional Data: A Kolmogorov-Smirnov Class Correlation-Based Filter. Recent Developments of Artificial Intelligence Methods, AI-METH Series 2009 2009
• W. Duch, T. Wieczorek, J. Biesiada, M. Blachnik Comparision of feature ranking methods based on information entropy.. In Proc. of International Joint Conference on Neural Networks. pp. 1415-1420. IEEE Press. Budapest, Hungary. 2004