lincgcg/Contextual-Outlier-Interpreter

Contextual-Outlier-Interpretation

This project provides an implementation for the paper:

Contextual Outlier Interpretation
Ninghao Liu, Donghwa Shin, Xia Hu
IJCAI 2018

Files in the folder

• data/
  • wbc/: a example dataset used in outlier detection
    • X.csv: each line represents one instance;
    • y.csv: labels indicating whether each instance is an outlier (1) or not (0);
• src/: implementations of the proposed outlier explanation method
  • main.py: runs the proposed method on the dataset
  • outlier_interpreter.py: implementation of the interpretation method
  • prediction_strength.py: estimates the number of clusters
  • utils.py: includes LOF and IsolationForest as outlier detectors