Hierarchical Dense Subtensor Detection in Tensors

CatchCore is a novel framework to detect hierarchical dense cores in multi-aspect data (i.e. tensors). CatchCore has the following properties:

• unified metric: provides a gradient-based optimized framework as well as theoretical guarantees
• accurate: provides high accuracy in both synthetic and real data
• effectiveness: spots anomaly patterns and hierarchical dense community
• scalable: scales almost linearly with all factors of input tensor, also has linearly space complexity

---

Datasets

The download links for the datasets used in the paper are available online.

• Android App rating. 1.32M × 61.3K × 1.28K × 5
• BeerAdvocate rating. 26.5K × 50.8K × 1472 × 1
• StackOverflow favorite. 545K × 96.7K × 1.15K × 1
• DBLP Co-author. 1.31M × 1.31M × 72
• Youtube Favorite. 3.22M × 3.22M × 203
• DARPA TCP Dumps. 9.48K × 23.4K × 46.6K
• AirForce TCP Dumps. 3 × 70 × 11 × 7.20K × 21.5K × 512 × 512

Environment

To install required libraries, please type

pip install -r requirements

---

Building and Running CatchCore

Please see User Guide

---

Running Demo

Demo for detecting hierarchical dense subtensor, please type

make

Reference

If you use this code as part of any published research, please acknowledge the following papers.

@article{feng2023hierarchical,
  title={Hierarchical Dense Pattern Detection in Tensors},
  author={Feng, Wenjie and Liu, Shenghua and Cheng, Xueqi},
  journal={ACM Transactions on Knowledge Discovery from Data},
  volume={17},
  number={6},
  pages={1--29},
  year={2023},
  publisher={ACM New York, NY}
}

@inproceedings{feng2019catchcore,
  title={CatchCore: Catching Hierarchical Dense Subtensor},
  author={Wenjie Feng, Shenghua Liu, and Xueqi Cheng},
  booktitle={European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD)},
  year={2019},
}