Deep anomaly detection with deviation network implementation (KDD19) by Guansong Pang, Chunhua Shen, Anton van den Hengel, University of Adelaide, Adelaide, Australia
The full paper can be found at https://dl.acm.org/citation.cfm?id=3330871 or https://arxiv.org/abs/1911.08623
This fork was refactored and utilizes Kaggle credit card fraud dataset available here: https://www.kaggle.com/mlg-ulb/creditcardfraud
- Download Kaggle dataset and place
csv
file indataset/
directory - Create
Python3.6
virtual env, activate it and runpip intall -r requirements.txt
- Run
devnet.py
and enjoy SOTA performance on Kaggle's Credit Card Fraud Dataset ๐ฉ