SFGAD is a tool for detecting anomalies in graph and graph streams with python.
I provides:
- Efficient computation of graph features
- Statistical models for detecting anomalous behavior
- Graph scanning to detect connected graph anomalies
- A customizable detection framework with 6 components
- Several pre-defined configurations
- Python: 3.5 or higher
- NumPy: 1.8.2 or higher
- SciPy: 0.13.3 or higher
- Pandas: 0.22.0 or higher
- NetworkX: 1.11.0 or higher
Installation of the latest release is available at the Python package index.
pip install sfgadThe source code is currently available on GitHub: https://github.com/sudrich/sf-gad
For testing use pytest from the source directory:
pytest sfgadThe framework defines an modular interface that allows full customization of the analysis process. For examples, see the tutorials on using a pre-defined analyzer and using a custom analyzer.
This work originated from the QuestMiner project (grant no. 01IS12051) and was partially funded by the German Federal Ministry of Education and Research (BMBF). The work was supported by the SDIL, which also maintains, this code base within the BMBF SDI-X Project (grant no. 01IS15035A)