A Hot Spot Detection Method of Criminal Events Under Data Uncertainty

Objectives Accurate crime hot spot detection technology plays an important role in decision-making support for urban public security management and police force deployment. However, due to the contingency and concealment of urban crime, criminal behavior is usually unable to be observed directly and accurately, resulting in the uncertainty of recorded crime information in time and space. Most of the existing hot spot detection methods are based on the assumption of complete data and they may get unreliable hot spot detection results when applied to practical crime datasets.

Methods Based on this, this paper proposes a crime hot spot detection method under data uncertainty. In terms of time, based on the idea of probability expression and Aoristic analysis, the time sequence law of crime is modeled. In terms of space, based on rational choice theory and daily activity theory, geographic weighted regression is used to correct the incomplete prior probability. Finally, the crime hot spots are extracted by expectation-maximization algorithm to draw a credible hot spot distribution map.

Results The experimental verification is based on the policing alert data of a prefecture level city in China. The results show that the proposed method can detect stable and reasonable hot spot results in the face of criminal events with different uncertainty strengths.

Conclusions The hot spots detected by the proposed method sufficiently eliminate the impact of the uncertainty of criminal events, fit with the daily activity theory in criminology, and can provide accurate decision support for police prevention and control deployment.