Sun, J., Yue, M., Lin, Z., Yang, X., Nocera, L., Kahn, G. et al. (2021). Crimeforecaster: Crime Prediction by Exploiting the Geographical Neighborhoods’ Spatiotemporal Dependencies. In Y. Dong, G. Ifrim, D. Mladenić, C. Saunders, & S. Van Hoecke (Eds.), Machine Learning and Knowledge Discovery in Databases. Applied Data Science and Demo Track (pp. 52-67). Springer International Publishing.
https://doi.org/10.1007/978-3-030-67670-4_4

has been cited by the following article:

• TITLE: Crime Prediction Based on Multi Perspective Feature Fusion and Extraction

  AUTHORS: Min Yu

  KEYWORDS: Deep Learning, Crime Prediction, Multi-View, GNNS

  JOURNAL NAME: Open Journal of Social Sciences, Vol.12 No.10, October 28, 2024

  ABSTRACT: The threat posed by criminal activities to the personal and property security of citizens, as well as to the harmony and stability of society, cannot be underestimated. Preventing crime has become one of the challenges in social governance. Crime prediction technology can extract potential crime patterns from historical crime data and predict future crime situations. The results of prediction can provide data support for formulating anti-crime policies and optimizing police patrol routes. One of the current challenges in the field of crime prediction lies in the independent extraction of features from the three most important perspectives: crime time, crime space, and crime type. Ignoring the other two perspectives while processing one may lead to the loss of useful information. This paper proposes a multi-perspective feature extraction and fusion technique, breaking the sequence of feature extraction among the three perspectives of time, space, and type. Based on the type node and bridging time and space, a multi-perspective fusion graph is constructed. The proposed method is thoroughly experimented on a real-world crime dataset. Experimental results demonstrate that the multi-perspective feature extraction and fusion technique proposed in this paper can achieve up to a 11.1% and 5.1% improvement in the Macro-F1 metric compared to the more recent crime prediction models Mist and Crime Forecaster, respectively.