系列案犯罪地理目标模型优化

Optimization of Criminal Geographic Targeting Model of Crimes in Series Cases

针对犯罪地理目标模型（CGT模型）在系列案件嫌疑人落脚点预测中未考虑地理环境因素影响,预测精度不高的问题,提出了一种顾及地理环境因素的犯罪地理目标模型优化方法（GEO-CGT模型）。研究采用相关性分析与灰色关联分析,刻画嫌疑人落脚点的地理环境相关性;参考多分类器系统理论,将地理环境因素与CGT模型进行非线性组合优化,并从搜索距离、面积误差对预测结果进行精度评估。以清远和韶关两市系列财产犯罪案件为样例数据,对模型预测进行对比实验,结果表明,改进后模型的预测精度相比于CGT和GEO-CGT模型均有显著提高。研究拓展了系列案嫌疑人落脚点预测方法,有效地提高了预测精度,对于警方缩小搜索范围,增大成功抓捕犯罪嫌疑人概率具有重要应用意义。

In light of the problem that the criminal geographic targeting（CGT） model fails to consider the impact of geographical environment factors and has low prediction accuracy in the prediction of the footholds of suspects in a series of cases, this article proposes a method of optimizing the criminal geographic targeting model taking into account of geo-environment. By using the correlation analysis and gray relational analysis,the research depicts the geo-environment relatedness of footholds of suspects; based on the multi-classifier system theory, conducts nonlinear combination optimization of the geo-environmental factors and the CGT model, and carries out accuracy assessment of predicting outcomes in terms of the search distance and the area error. Taking the series of property crime cases in Qingyuan and Shaoguan as sample data, a comparative experiment on model prediction shows that the prediction accuracy of the improved model is significantly higher than that of CGT and GEO-CGT models. The research expands the method of predicting the footholds of suspects in a series of cases and effectively improves the prediction accuracy, demonstrating a great significance for the police to narrow the search scope and increase the probability of successful arresting of suspects.