Crime early warning method, device, computer equipment and storage medium

Abstract

The embodiment of the invention discloses a crime early warning method, device, computer equipment and storage medium. The method includes: acquiring case data; inputting the case data into a multi-modal crime early warning model for crime identification to obtain identification results; wherein, the multi-modal crime early warning model pre-warns several multi-source data The sample set formed after processing and grouping is obtained by training a multimodal deep learning network. By implementing the method of the embodiment of the present invention, the potential information of multi-source heterogeneous data can be fully utilized to actively mine effective features, and the accuracy of crime early warning can be improved.

Description

technical field [0001] The present invention relates to the technical field of crime early warning, and more particularly to a crime early warning method, device, computer equipment and storage medium. Background technique [0002] With the rapid development of society, the manifestations of various crimes are also becoming more and more diverse. If accurate early warning of criminal behavior can be carried out, prevention and control measures can be taken in advance. How to avoid and prevent the occurrence of criminal behavior is an urgent problem to be solved at this stage. Crime early warning refers to the process of using specific methods to mine crime patterns from a large amount of historical data, and give advance warnings before criminal behavior occurs. [0003] Data mining technology and artificial intelligence-related algorithms can quickly and efficiently mine crime characteristics from massive data and predict future behavior. The public security industry is a ...

AI Technical Summary

Problems solved by technology

[0004] Chinese patent 202010601748.2 only focuses on video streaming data, and does not make good use of multi-source data in historical crime data

Moreover, this method does not involve unstructured data modeling, and has not yet fully utilized the multi-source data information of public security and policing. It can be seen that most of the traditional multi-modal crime early warning models rely on the logical characteristics and correlations summarized by experience. Combined with specific numerical features to establish a traditional supervised analysis model, the data structure is single, the information obtained is limited, and the potential information of multi-source heterogeneous data is not fully utilized to actively mine effective features. The trained model is accurate in crime warning. not tall

Images