MLkid implements a small set of machine learning modules with Python and c++.

The sample data are mostly from the UCI Machine Learning Repository: http://archive.ics.uci.edu/ml/

Among all the packages, the jsvm and jmaxent are c++ implemented.

• jsvm elegantly implement the SMO algorithm, thus can be used for understanding and teaching SMO.
• jmaxent is an efficient package for maximum entropy model, providing three types of sovlers: L1-OWLQN, L1-SGD, L2-LBFGS. This is for practical use and performs quite well.

Other packages are mostly toy programs, just for fun and study.