Anomaly-Detection-in-Networks

Course Project of Compressed Sensing

Citation

This project is an implementation of the anomaly detection method proposed in the following paper:

@article{elliott2019anomaly,
  title={Anomaly Detection in Networks with Application to Financial Transaction Networks},
  author={Elliott, Andrew and Cucuringu, Mihai and Luaces, Milton Martinez and Reidy, Paul and Reinert, Gesine},
  journal={arXiv preprint arXiv:1901.00402},
  year={2019}
}

Prerequisite

1. You need to have networkx, rpy2 python packages installed.
2. Following the instruction of this github repository: https://github.com/taynaud/python-louvain to install python-louvain package which performs community detection.
3. Following the instruction of this github repository: https://github.com/alan-turing-institute/network-comparison to install netdist R package which performs network comparison. And rpy2 package is necessary to utilise R package in python.

Structure

1. Under the main directory, there is a main.py to generate data under the current folder.
2. We also prepared some generated data under the directory: ./data.
3. Under the main directory, there is a jupyter notebook: demo.ipynb, including a demo of data generation, and a demo of how to perform Anomaly Detection algorithm on the generated dataset.