This repository is the source code the paper: https://arxiv.org/abs/2206.06602
Please cite our paper if you use this repository.
@article{xu2022deep,
title={Deep Isolation Forest for Anomaly Detection},
author={Xu, Hongzuo and Pang, Guansong and Wang, Yijie and Wang, Yongjun},
journal={arXiv preprint arXiv:2206.06602},
year={2022}
}
All the experiment results reported in our paper can be well reproduced.
use python main.py --runs 10 --model dif to run our model DIF,
use python main_graph.py --runs 10 --model dif to perform the experiments on graph data
GLocalKD and InfoGraph can be directly used after downloading their implementation from Github.
use python main_ts.py --runs 10 --model dif to perform experiments on time-series data
GDN and Omni are also publicly aviable and can be directly used after downloading from Github
Five ablated varients DIF-AE, DIF-DSVDD, RR-COPOD, RR-KNN, and RR-LOF are supported in this project as well.
Please add --contamination 0.1 when performing main.py
The synthetic datasets can be created by create_scal_data.py.
we record the execution time in the final record files. After each running, a record file will be generated.
Change the --n_ensemble 50 (number of representations) and --n_estimators 6 (number of trees per representation) to other settings.
Different network structures are implemented as well, please change the argument --network_name
We retain two small tabular datasets (Ad and Cardio), one time-series dataset (MSL), and one graph dataset (MMP) as examples. For other datasets, we provide links in the appendix of our paper.
Ad and Cardio are obtained from https://archive.ics.uci.edu/
MSL is obtinaed from https://github.com/khundman/telemanom
MMP is obtained from https://chrsmrrs.github.io/datasets/docs/datasets/
Copyright of dataset MSL:
Copyright Assertion
Copyright (c) 2018, California Institute of Technology ("Caltech"). U.S. Government sponsorship acknowledged.
All rights reserved.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
• Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
• Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
• Neither the name of Caltech nor its operating division, the Jet Propulsion Laboratory, nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent