graph-fraud-detection-papers

Why we included this project

Fraud, anomaly, and outlier detection now spans both graph neural networks and transformer/LLM approaches, and this repository tracks both sides of that literature. The papers are grouped by year and by method, with a separate section for non-deep-learning work going back to 2014, and each entry links to the paper, venue, and code when available, so you can move from reading an abstract to reproducing a method without hunting through search engines. It also collects toolboxes, datasets, and surveys, which helps when you need a baseline and a benchmark for an experiment. The companion dashboard and a local RAG-based chatbot let you filter and query the same several hundred papers without downloading everything, which is handy for a broad literature review. If you are scoping prior work before building a detection model, this is a solid first stop.