Crime prediction method based on improved STARMA model

Abstract

The invention discloses a crime prediction method based on an improved STARMA model. The method comprises the following steps: (1) obtaining a taxi, a crime, a complaint telephone data set and topographic information of a to-be-analyzed area; (2) preprocessing the data set, and extracting a complaint telephone matrix according to the complaint telephone data set; converting the crime data into a stable sequence; (3) extracting a travel adjacent lag matrix according to the terrain information and the taxi data; (4) performing space-time correlation analysis on the stable crime data and the travel adjacent lag matrix, and determining the time and space delay order of the crime data; and (5) based on the time and space delay orders obtained in the step (4), combining the stable crime data obtained in the step (2) and the step (3), the travel adjacency lag matrix and the complaint telephone matrix to realize the crime prediction based on the improved STARMA model. The method is high in prediction precision.

Description

Technical field [0001] The invention relates to the field of crime prediction and data mining, in particular to a crime prediction method based on an improved STARMA model. Background technique [0002] In recent years, there has been a marked increase in various types of criminal cases in my country, and the pressure on public security departments has also increased, which has affected the stability of social order and the development of social economy. In the face of severe social situations, crime prediction can provide a theoretical basis for targeted police deployment to prevent crime. [0003] The distribution and occurrence of criminal acts have certain regularity in time and space. Therefore, on the one hand, predictive analysis can be performed from the perspective of time, such as time series method, double moving average method, exponential smoothing method, leading indicator method, factor modeling method, etc.; on the other hand, spatial sequence method can also be us...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the shortcomings of the existing crime prediction methods, such as low prediction accuracy and few features, the present invention proposes a crime prediction method based on the improved autoregressive moving average model (Time spatial autoregressive moving average, STARMA), which is different from the current crime prediction method , for the spatial weight matrix, a new matrix (travel adjacency lag matrix) construction method is proposed by combining static regional information and dynamic taxi flow information to measure the crime correlation between regions. The constructed travel adjacency lag matrix has the The time factor is different from the current static weight matrix construction method

Images