Crime space-time risk prediction and decision scheduling method and device

Abstract

The invention discloses a crime space-time risk prediction and decision scheduling method and device. The method comprises the following steps: S100, acquiring data resources and establishing a corresponding database; s200, performing data governance to obtain output data; s300, on the basis of a deep learning CTR estimation model and the output data, estimating the case issuing probability of the road section to be predicted in the next week; s400, based on the case issuing probability of the road section to be predicted in the next week, the receptive field filter, the jurisdiction filter, the hotspot migration rule base and the historical case event database, performing regularization to obtain the corrected case issuing probability of the road section to be predicted in the next week; and S500, police scheduling and patrol control route planning are carried out. According to the method, the input flexibility of the prediction model can be enhanced, the mobility of the system can be improved, the interpretability of the prediction result can be enhanced, the prediction result correction flexibility can be improved by adding the prediction probability correction module, and the research, judgment and scheduling process of the prediction result can be simplified by adding the scheduling module.

Description

technical field [0001] The invention relates to the field of spatio-temporal information processing, in particular to a crime spatio-temporal risk prediction and decision scheduling method and device. Background technique [0002] Today's mainstream crime space-time risk prediction system mainly includes two modules: data governance and future crime probability prediction. Among them, the data governance module will process the input multi-source heterogeneous data into a format that the probability estimation module can input. After receiving the processed data, the future probability prediction module conducts online or offline training. After the prediction model reaches a suitable accuracy, it goes online and provides forecasting services. The parameters of the online model can be updated in real time according to the data. The mode of offline update can also be adopted to adapt to the change of data. [0003] In the prior art, the mainstream technology of crime spatio...

AI Technical Summary

Problems solved by technology

[0008] The prediction model in the existing prediction system has poor portability. For the grid-based model, the size of the input data dimension changes due to the change of the predicted spatial range, or the input data source changes due to the change of the predicted spatial location. etc., it is possible to redesign and train the already trained model

[0009] The existing forecasting system does not have a reasonable interpretability of the forecasting results, and cannot correct the forecasting errors caused by the deviation of the forecasting model itself

Most of the existing prediction models are based on deep learning models. Although the prediction accuracy is generally not too low after a large amount of data training, there is no system to judge whether the prediction results are reasonable when the prediction results are obtained, not only cannot explain the data The mechanism behind it cannot correct the erroneous results caused by the model's own deviation

Images