In this release we bring again quite some improvements to the library. Most important to note is that we have modified the structure of the feature selection package to better reflect the classification of feature selection algorithms and we have added some feature selection algorithms.
- Improvements to a number of algorithms with some extra utility constructors to make them easier to use.
Extra regressions tests to cover more code.
Improved the documentation of several algorithms
Reorganized the structure of the feature selection package to a more logical structure.
Added greedy forward selection and backward elimination algorithms for feature selection
Parametrized the k-nearest neighbors algorithm to be able to use any distance metric.
Added more tutorials