Reproduce “HAGEN: Homophily-Aware Graph Convolutional Recurrent Network for Crime Forecasting” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Chenyu Wang, Zongyu Lin, Xiaochen Yang, Jiao Sun, Mingxuan Yue, Cyrus Shahabi

HAGEN forecasts next-slot crime categories for regions in Chicago and Los Angeles. It learns a directed graph from crime data, encouraging linked regions to share crime labels, then uses graph diffusion and historical records to make predictions. Across five test months, it outperforms CrimeForecaster on Micro-F1 and Macro-F1. The results do not establish causal crime flows or improved deployment outcomes.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.