系列案犯罪地理目标模型优化

doi:10.13249/j.cnki.sgs.2018.08.002

摘要/Abstract

摘要：

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。

关键词: 系列犯罪案件, 嫌疑人落脚点, 地理环境因素, CGT模型

Abstract:

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.

Key words: serial criminal cases, suspects′ location;, geographical environment factors, CGT model

中图分类号:

P208

方嘉良, 李卫红. 系列案犯罪地理目标模型优化[J]. 地理科学, 2018, 38(8): 1210-1217.

Jialiang Fang, Weihong Li. Optimization of Criminal Geographic Targeting Model of Crimes in Series Cases[J]. SCIENTIA GEOGRAPHICA SINICA, 2018, 38(8): 1210-1217.

图/表 7

表1

表2

表3

表4

表5

图1

图2

参考文献 30

[1]	Cohen L E, Felson M.Social change and crime rate trends: A routine activity approach[J]. American Sociological Review, 1979: 588-608.
[2]	Cornish D B, Clarke R V.The reasoning criminal: Rational choice perspectives on offending[M]. Transaction Publishers, 2014.
[3]	Brantingham P L, Brantingham P J.Patterns in crime[M]. New York: Macmillan, 1984.
[4]	孙峰华, 毛爱华. 犯罪地理学的理论研究[J].人文地理, 2003, 18(5):70-74.
	[Sun Fenghua, Mao Aihua.Study on the theories of criminal geography. Human Geography, 2003, 18(5):70-74.]
[5]	Gorman D M, Speer P W, Gruenewald P J.Spatial dynamics of alcohol availability, neighborhood structure and violent crime[J]. Journal of Studies Alcohol and Drugs, 2001, 62(5): 628-636. doi: 10.15288/jsa.2001.62.628
[6]	Ceccato V, Haining R, Kahn T.The geography of homicide in Sao Paulo, Brazil[J]. Environment & Planning A, 2007, 39(7):1632-1653.
[7]	Ajimotokin S, Haskins A, Wade Z.The effects of unemployment on crime rates in the U.S.[J]. Journal of Biotechnology, 2015, 186:98-109.
[8]	Alwee Razana, Shamsuddin Siti Mariyam H, Sallehuddin Roselina Salleh. Economic indicators selection for crime rates forecasting using cooperative feature selection[A]. 20th National Symposium on Mathematical Sciencess[C]. 2013.
[9]	毛媛媛, 戴慎志. 犯罪空间分布与环境特征——以上海市为例[J]. 城市规划学刊, 2006(3):85-93.
	[Mao Yuanyuan, Dai Shenzhi.Research on spatial and environmental characters of crimes: case study of Shanghai. Urban Planning Forum, 2006(3):85-93.]
[10]	颜峻, 疏学明, 袁宏永.盗窃犯罪空间分布与地理因素的关联[J]. 清华大学学报(自然科学版), 2010(2):174-176.
	[Yan Jun, Shu Xueming, Yuan Hongyong.Relationship between spatial distribution of thief crime and geographical factors. Journal of Tsinghua University, 2010(2):174-176.]
[11]	柳林, 张春霞, 冯嘉欣, 等. ZG市诈骗犯罪的时空分布与影响因素[J]. 地理学报, 2017, 72(2):315-328.
	[Liu Lin, Zhang Chunxia, Feng Jiaxin et al. The spatial-temporal distribution and influencing factors of fraud crime in ZG city, China. Acta Geographica Sinica, 2017, 72(2):315-328.]
[12]	杨刚斌, 柳林, 何深静, 等. 基于多层次模型的广州门禁小区家庭入室盗窃环境影响分析[J]. 热带地理, 2016, 36(4):610-618. doi: 10.13284/j.cnki.rddl.002860
	[Yang Gangbin, Liu Lin, He Shenjing et al. Environmental impacts on burglary victimization in gated communities: a multi-level analysis in Guangzhou. Tropical Geography, 2016, 36(4):610-618.] doi: 10.13284/j.cnki.rddl.002860
[13]	周素红, 谢蔚翰, 宋广文, 等. 土地利用对街头抢劫影响的空间分异模式: 以H市为例[J]. 地理科学, 2017, 37(6):885-894.
	[Zhou Suhong, Xie Weihan, Song Guangwen et al. The spatial differentiation effect of land use on street robbery:a case study in H City, China. Scientia Geographica Sinica, 2017, 37(6):885-894.]
[14]	陈鹏, 李欣, 胡啸峰, 等. 北京市长安街沿线的扒窃案件高发区分析及防控对策[J]. 地理科学进展, 2015, 34(10):1250-1258.
	[Chen Peng, Li Xin, Hu Xiaofeng et al. Clustering pattern analysis and prevention strategies to pickpocketing offence along the Chang'an Street in Beijing. Progress in Geography, 2015, 34(10):1250-1258.]
[15]	郑文升, 卓蓉蓉, 罗静, 等. 基于空间句法的武汉城区“两抢一盗”犯罪分布环境[J]. 地理学报, 2016, 71(10):1710-1720.
	[Zheng Wensheng, Zhuo Rongrong, Luo Jing et al. The distribution environment of robbery, snatch and theft crime based on space syntax: A case study of the central area of Wuhan. Acta Geographica Sinica, 2016, 71(10):1710-1720.]
[16]	宋广文, 肖露子, 周素红, 等. 居民日常活动对扒窃警情时空格局的影响[J]. 地理学报, 2017, 72(2):356-367.
	[Song Guangwen, Xiao Luzi, Zhou Suhong et al. Impact of residents' routine activities on the spatial-temporal pattern of theft from person. Acta Geographica Sinica, 2017, 72(2):356-367.]
[17]	齐杰, 王国梁. 太原市城市发展对城市问题的影响[J]. 城市地理, 2017(8):7-8.
	[Qi Jie, Wang Guoliang.Impact of urban development on urban problems in Taiyuan. Cultural Geography, 2017(8):7-8.]
[18]	冯健, 黄琳珊, 董颖, 等. 城市犯罪时空特征与机制——以北京城八区财产类犯罪为例[J]. 地理学报, 2012, 67(12):1645-1657.
	[Feng Jian, Huang Linshan, Dong Ying et al. Research on the spatial-temporal characteristics and mechanism of urban crime: a case study of property crime in Beijing. Acta Geographica Sinica, 2012, 67(12):1645-1657.]
[19]	王占宏. 基于扫描统计方法的上海犯罪时空热点分析[D]. 上海: 华东师范大学, 2013.
	[Wang Zhanhong.Hotspot analysis of crime based on scanning statistical method in Shanghai. Shanghai: East China Normal University, 2013.]
[20]	郑滋椀. 基于道路网络的犯罪时空分布特征与可视化研究[D]. 杭州: 浙江大学, 2016.
	[Zheng Ziwan.Temporal and spatial distribution characteristics and visualization of crime based on road network. Hangzhou: Zhejiang University, 2016.]
[21]	毛媛媛, 丁家骏. 抢劫与抢夺犯罪行为时空分布特征研究——以上海市浦东新区为例[J]. 人文地理, 2014(1):49-54.
	[Mao Yuanyuan, Deng Jiajun.Study on spatial-temporal patterns of robbery and snatch: a case study of pudong new area, Shanghai. Human Geography, 2014(1):49-54.]
[22]	Rossmo D K.Geographic profiling: Target patterns of serial murderers[D]. Burnaby: Theses (School of Criminology)/Simon Fraser University, 1995.
[23]	Canter D, Larkin P.The environmental range of serial rapists[J]. Journal of Environmental Psychology,1993,13(1): 63-69. doi: 10.1016/S0272-4944(05)80215-4
[24]	吴沂霖. 连环犯罪案件中地点和时间的预测模型[J]. 无线互联科技, 2015(13):120-123.
	[Wu Yilin.Forecast model of location and time in the case of a serial crime. Wireless Internet Technology, 2015(13):120-123.]
[25]	章社生, 曲径, 郑传建, 等. 犯罪嫌疑人住处及下次作案地点预测[J].武汉理工大学学报, 2010, 32(6):970-973.
	[Zhang Shesheng, Qu Jing, Zheng Chuanjian et al. Prediction of the residence of the criminal and the location of the next crime. Journal of Wuhan University of Technology, 2010, 32(6):970-973.]
[26]	徐冲, 柳林, 周素红. 基于临近相似性考虑的犯罪热点密度图预测准确性比较——以DP半岛街头抢劫犯罪为例[J]. 地理科学, 2016, 36(1):55-62. doi: 10.1329/j.cnki.sgs.2016.01.007
	[Xu Chong, Liu Lin, Zhou Suhong.The comparison of predictive accuracy of crime hotspot density maps with the consideration of the near similarity: a case study of robberies at DP peninsula. Scientia Geographica Sinica, 2016, 36(1):55-62.] doi: 10.1329/j.cnki.sgs.2016.01.007
[27]	肖汉, 杜永慧, 徐金泽, 等. 融合GIS的犯罪概率模型及应用[J]. 北京大学学报(自然科学版), 2013, 49(6):1025-1030.
	[Xiao Han, Du Yonghui, Xu Jinze et al. Integration of the GIS with criminal probability model and its application. Acta Scientiarum Naturalium Universitatis Pekinensis, 2013, 49(6):1025-1030.]
[28]	李新光, 章家岩, 冯旭刚, 等. 基于模糊算法的犯罪地理目标模型优化[J]. 常州工学院学报, 2017, 30(4):28-32.
	[Li Xinguang, Zhang Jiayan, Feng Xugang et al. An optimization of geographic profiling model based on fuzzy algorithm. Journal of Changzhou Institute of Technology, 2017, 30(4):28-32.]
[29]	李卫红, 戴侃, 闻磊. 顾及地理因素的犯罪地理目标模型改进方法[J]. 测绘科学, 2015, 40(7):86-91.
	[Li Weihong, Dai Kan, Wen Lei.Improvement of criminal geographic targeting model considering geographic factors. Science of Surveying and Mapping, 2015, 40(7):86-91.]
[30]	付彬. 多分类器组合中的基分类器选取方法[D]. 北京: 北京交通大学, 2009.
	[Fu Bin.Classifier selection method in multi-classifier combination. Beijing: Beijing Jiaotong University, 2009.]

类型	因子	单位
警方排查力度	案件数	个
	岗亭	个
	监控点	个
地区发达情况	路网	m
土地利用类型	出租屋	个
	娱乐场所	个
	裸地	m²
	交通用地	m²
	植被	m²
	居民地	m²
	商服用地	m²
	机关组织用地	m²
	耕地	m²
	水体	m²
	医疗卫生用地	m²
	公共用地	m²

	案件数	出租屋	岗亭	监控点	娱乐场所	路网	居民地	植被
嫌疑人落脚点	0.271^**	0.224^**	0.082^*	0.203^**	0.012^**	0.002^**	0.002	0.005
	裸地	交通用地	商服用地	机关用地	耕地	水体	医疗	公共
嫌疑人落脚点	0.075^**	0.106^**	0.062^**	0.032^**	-0.022^*	-0.005	0.084^**	0.043

	案件数	出租屋	岗亭	监控点	娱乐场所	路网	居民地	植被
嫌疑人落脚点	0.419^**	0.419^**	0.122^*	-0.468^**	0.320^**	0.219^**	0.265^**	0.013
	裸地	交通用地	商服用地	机关用地	耕地	水体	医疗	公共
嫌疑人落脚点	0.049^**	0.186^**	0.401^**	0.314^**	-0.121^*	-0.036	0.193^**	0.072

	案件数	出租屋	岗亭	监控点	娱乐场所	路网	居民地	植被
嫌疑人落脚点	0.903	0.856	0.888	0.843	0.783	0.726	0.816	0.756
	裸地	交通用地	商服用地	机关用地	耕地	水体	医疗	公共
嫌疑人落脚点	0.723	0.856	0.806	0.744	0.643	0.754	0.793	0.802

	案件数	出租屋	岗亭	监控点	娱乐场所	路网	居民地	植被
嫌疑人落脚点	0.928	0.966	0.966	0.928	0.967	0.909	0.906	0.906
	裸地	交通用地	商服用地	机关用地	耕地	水体	医疗	公共
嫌疑人落脚点	0.824	0.911	0.912	0.856	0.734	0.856	0.876	0.854