Scikit-learn Greedy Maximum Deviation (GMD) Algorithm

This project provides a scikit-learn compatible python implementation of the algorithm presented in [Trittenbach2018] together with some usage examples and a reproduction of the results from the paper.

Recent approaches in outlier detection seperate the subspace search from the actual outlier detection and run the outlier detection algorithm on a projection of the original feature space. See [Keller2012]. As a result the detection algorithm (Local Outlier Factor is used in the paper) does not suffer from the curse of dimensionality.

Refer to the documentation to see usage examples.