QQ客服 官方微博 反馈留言

题名	警务情报数据分析方法研究及应用
作者	张素洁
学位类别	硕士
答辩日期	2017-05-24
授予单位	中国科学院沈阳自动化研究所
授予地点	沈阳
导师	赵怀慈
关键词	警情数据 时间序列化 聚类 关联规则
其他题名	Research and Application of Police Intelligence Data Analysis Method
学位专业	控制工程
中文摘要	本论文对警情数据进行分析就是在这一背景下建立起来的，从表面上看杂乱无章、毫无关联的大量报警数据中找出有用的信息，进而提高公安部门情报分析能力。本论文的研究内容主要包括以下四个方面。警情数据的时间序列化分析。对训练集中的警情数据进行时间序列化分析，采用合适的模型对警情数据的月序数列进行建模，用测试集中的警情数据验证模型拟合是否具有较高的准确率。再对未来犯罪量进行定量预测，预测未来六个月警情的趋势走向，对社会治安形势的判断和治理政策的调整提供科学性的指导。基于聚类分析的案件模式研究。对警情数据进行k-means聚类分析，在分析的过程中选取合适的k值，基于聚类中心点所在的周围区域相对比较密集，并且聚类中心点之间距离相对较远的选取原则，找出最优的k个初始聚类中心点。通过可视化每一类别警情的分布模式特点，找出每一类别的热点区域、热点时间，合理的调度警力部署。警情数据中关联规则的挖掘。传统的Apriori算法在进行频繁项集挖掘的过程中需要重复扫描数据库，并且剪枝的过程中会产生大量候选频繁集，针对这两个不足，提出了基于Boolean矩阵对Apriori算法进行改进。采用改进后的算法对聚类后各个类别中的警情数据进行关联规则的挖掘，找出警情的地点、时间、类型和天气这四个属性之间潜在的联系，以此来预测案件发生的规律。最后，对警情数据分析系统进行了设计与实现。基于JavaEE平台MVC设计模式对系统进行三层架构设计，层次清晰逻辑性强。为将来进一步研究提供了基本的软件支持。 本课题以深圳市龙华公安分局三年来提供的警情数据为基础，研究了深圳市龙华公安分局历史警情趋势和警情的特点以及各警情属性的关联规则，为深圳市龙华公安分局打击和预防犯罪提供科学的参考依据。
英文摘要	This paper analyzes the police intelligence data is established in this context from a large number of chaotic, unrelated data to find useful information to improve the ability of public security intelligence analysis. The main contents of this thesis include the following four aspects. The time series analysis of the alarm data in the training set is carried out. The appropriate model is used to model the monthly sequence of the alarm data, Test whether the model fit has a high accuracy by using the test data .And then predict the trend of the next six months police intelligence to provide a scientific guide judgment of social security situation and the adjustment of governance policy. Research on case model based on cluster analysis. The k-means clustering analysis of the alarm data is carried out. In the process of analysis, select the appropriate k value, based on the clustering center point where the surrounding area is relatively dense, and the distance between the clustering center distance is relatively far from the selection principle, select the optimal k initial clustering center point. By visualizing the distribution pattern of each category of police intelligence, we can find out the hotspot area, hot time and reasonable dispatching of each category. Excavation of association rules in police intelligence data. The traditional Apriori algorithm requires repeated scanning of the database in the process of frequent item sets mining, and the pruning process will produce a large number of candidate frequent sets. In view of these two problems, the Apriori algorithm is improved based on the Boolean matrix. The improved algorithm is dig the association rules in each category after clustering, to find out the potential link between the four attributes of location, time, type and weather. Finally, the police intelligence data analysis system was designed and implemented. Based on the JavaEE platform MVC design pattern, the system is designed in three layers, and the level is clear and logical. The software provides a basic foundation condition for further research. Based on the police data provided by the Shenzhen Longhua Public Security Bureau in the past three years, this paper studies the trend of the historical police and the characteristics and the rules of the association of police intelligence, providing scientific reference for Shenzhen Longhua Public Security Bureau to combat and prevent crime.
语种	中文
产权排序	1
内容类型	学位论文
源URL	[http://ir.sia.cn/handle/173321/20542]
专题	沈阳自动化研究所_光电信息技术研究室
推荐引用方式 GB/T 7714	张素洁. 警务情报数据分析方法研究及应用[D]. 沈阳. 中国科学院沈阳自动化研究所. 2017.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。