连环作案分析.docxVIP

对于已经锁定在一个地区内作案的连环作案分析利用预测罪犯的下一犯罪地点。对于一个区域内的连环作案分析，我们根据作案人的思维定势总结出罪犯作案的特点，（由于主要是判断罪犯的下一犯罪地点，我们要详细讨论其作案的地理属性特点：地形，人口密度，设施等）。另一方面，在一般的刑事案件中，警察破案时，会在大量信息的筛选排查中锁定一些范围，在锁定的范围里进行蹲点守候。对于某一地区，由于警察局破案人员是有限的。因此，在有限的警力资源下，要想尽快找到罪犯的下一犯罪地点就必须在筛选蹲守地点时寻找最合适的地点.因此，这一部分的模型如下：1 建立警方警力部署数据库：对于当地警局，警局实现根据在实际侦查中总结的经验，将全市（假设锁定该区域为市）按照警察实际实行监控能力（平均标准，不考虑警察能力差异）进行区域划分（可根据街道，以及执行能）。每一个区域内可以有不同的设施，地形，但是必须是一组警力可以控制的范围。有些区域划分不到怎么办？只要是该区域内有人居住或经常性活动，则该区域应该划分到警力控制区域内，如果某些个别地区如山林，水库等地人烟稀少，则可以忽略不划分警力范围，但在数据库建设说明中要说明（防止遗漏）有些区域划分重复？为了避免警力浪费1 利用数学方法对连环案件中警方掌握的各种有效信息进行统计分析。将每一次案件中的各种信息以表格形式罗列出来：属性案件时间地点人口密集度……杀人杀人杀人1Introduction2 Assumption3 Scheme 1:Region matchingmodel based on grid1.1 General partition of a regionTo begin with the model of predicting possible locations of the next crime, we try to divide the regioninto different areas, as residential districts, downtown, public open space, college and school areas and industrial zones. More specifically each area is divided into different districts, each of whichcould be totally controlled by one group of police forces. As we assumed above that the police forces are never sufficient to oversee every corner of the city, comparing with the uncertainty of the next possible criminal location, this kind of division is more feasible to monitor one district or assign police there. After the partition, it is helpful to mark these districts with serial numbers, i.e. representsevery district in the region.Classification of attributes1.2.1Summarizing the relative attributesof serial crimes’ locationsFocusing on the kind of the serial crime we are investigating, police could abstract some crucial qualities of this crime and exclude many inconsequential ones, according to the officers’previous experience and the possible relating attributes given by the inference engine that has compiled the analogous files and has been trained to summarize the likely correlative qualities of this type of cases.These key attributes could be divided into two types:2-valuedattr