摘要
针对犯罪地理目标模型(CGT模型)在嫌疑人落脚点预测中没有考虑落脚点选择的地理相关性和部分落脚点不在预测范围内两方面的不足,该文提出了顾及地理环境因素的犯罪地理目标改进模型(GEO-CGT模型):基于圆圈假设将模型计算区域扩展,解决部分落脚点不在预测范围问题;利用空间相关性分析,确定影响嫌疑人落脚点选择的地理因素,建立GEO-CGT模型;利用已有案件数据对模型进行对比验证和精度评估。结果表明,地理环境因素对嫌疑人落脚点选择具有较大的影响,改进后的GEO-CGT模型能有效减少搜寻排查嫌疑人的距离与面积,提高预测精度。研究成果可为同类研究提供参考。
Aiming at the insufficiency of Criminal Geographic Targeting(CGT) models in predicting crime suspects' hideouts that the roles of geographical factors in choosing the hideouts were not considered, and some hideouts lay out of the predicted ranges, the paper proposed the improved CGT (GEO- CGT) model which takes into account geographical factors: a circle hypothesis was applied to extend the calculation areas of the model to solve the problem that some hideouts lay out of the predicted ranges; the correlation analysis was carried out to determine the geographical factors that influenced the suspects' selection of their hideouts and help design the GEO-CGT model; and real crime data was used to assess the effects of the improved model. Result showed that the geographical factors would have significant effects on suspects' selection of their hideouts, and the improved CGT model may effectively reduce the distance and area in searching suspects and enhance the prediction accuracy, which could provide a reference for related studies.
出处
《测绘科学》
CSCD
北大核心
2015年第7期86-91,149,共7页
Science of Surveying and Mapping
关键词
地理画像
犯罪地理目标模型
犯罪活动距离
嫌疑人落脚点
预测
geographic profiling
Criminal Geographic Targeting(CGT) model
journey-to-crime distance
suspect hideout
prediction