Uncertainty-Aware Crime Prediction With Spatial Temporal Multivariate Graph Neural Networks

TL;DR

This paper introduces STMGNN-ZINB, a novel graph neural network framework that effectively models sparse, non-Gaussian crime data by combining diffusion, convolution, and zero-inflated negative binomial distributions for improved urban crime prediction.

Contribution

The paper presents a new spatial-temporal multivariate GNN model that handles data sparsity and non-Gaussian distributions in crime forecasting, outperforming existing methods.

Findings

STMGNN-ZINB achieves higher prediction accuracy than baseline models.

The model provides more reliable confidence intervals for crime predictions.

Evaluation on real datasets confirms its effectiveness in urban crime analysis.

Abstract

Crime forecasting is a critical component of urban analysis and essential for stabilizing society today. Unlike other time series forecasting problems, crime incidents are sparse, particularly in small regions and within specific time periods. Traditional spatial-temporal deep learning models often struggle with this sparsity, as they typically cannot effectively handle the non-Gaussian nature of crime data, which is characterized by numerous zeros and over-dispersed patterns. To address these challenges, we introduce a novel approach termed Spatial Temporal Multivariate Zero-Inflated Negative Binomial Graph Neural Networks (STMGNN-ZINB). This framework leverages diffusion and convolution networks to analyze spatial, temporal, and multivariate correlations, enabling the parameterization of probabilistic distributions of crime incidents. By incorporating a Zero-Inflated Negative Binomial model, STMGNN-ZINB effectively manages the sparse nature of crime data, enhancing prediction accuracy and the precision of confidence intervals. Our evaluation on real-world datasets confirms that STMGNN-ZINB outperforms existing models, providing a more reliable tool for predicting and understanding crime dynamics.