V0.5.0

In this release, more algorithms are implemented, including PCA and Feature Bagging models. The documentation and test coverage are also improved. A comparison for PyOD algorithms is presented below:

v0.4.5

This is the first Github release to sync with PyPI. The current version brings the following changes compared with v0.3.x:

• Supported both Python 2 and 3 now.
• Refactor all models to be sklearn compatible. The models can be easily cloned, modified and viewed.
• Simplified examples for easy reproduction.
• Optimized code for HBOS and ABOD. Codes are refactored for better stability.
• Added more tests to improve the coverage.
• Added two combination functions, Average() and Maximization()