oddball_py3

Python3 implementation of oddball

This is the Python3 implementation of oddball.
For more detail, you can see: Akoglu L., McGlohon M., Faloutsos C.(2010) oddBall: Spotting Anomalies in Weighted Graphs.

Environments

networkx (version: 2.1)
numpy
scikit_learn
You can use the following command to download the environments directly.

pip install -r requirements.txt

Run

The input is a weighted undirected graph which format is 'edge1 edge2 weight'.

Options:

--input: input file
--output: output file
--lof: Use LOF. 0: not use. 1: use. Default value is 0.
--anomaly_type: Anomaly Type. 1:star_or_clique. 2:heavy_vicinity. 3:dominant_edge.

You can use --help for more details.
Here is a sample.

python main.py --input inputFile --output outputFile --lof 0 --anomaly_type 1

Miscellaneous

Hope you will like this repository and have fun.
Don't forget a star, lol~