基于CEP并行化处理的犯罪线索实时推荐研究
发布时间:2019-05-19 00:02
【摘要】:纵观历史,犯罪从国家产生之日就与之相伴随,存在于各个历史阶段和各种社会类型,从未消失。网络技术的诞生和发展,为案件侦查提供更加广阔的途径。依靠网络开展犯罪线索搜集将成为案件侦查的新方法。随着互联网技术的高速发展,网络的信息数据已经呈指数增长,因此如何从海量的数据中快速的查找有价值的线索并实时地推荐给相关部门是案件侦查亟需解决的问题。复杂事件处理(Complex Event Processing,CEP)是将外部源源不断的数据流抽象为事件,再通过对事件的过滤、聚合等操作,检测事先定义好的复杂事件模式并对其进行响应的过程。它最大的优势是处理实时数据,这能让我们在最短的时间内获取最有用的信息。复杂事件处理的并行化有效的解决了集中式处理存在的性能问题,在面对海量数据时表现出极大的优势。在原有犯罪线索收集的基础上,引入并行化的复杂事件处理技术,为相关部门推荐犯罪线索提供实时性和高效性的保证。本文通过对复杂事件处理并行化技术进行研究,在现有的网络犯罪线索收集系统的基础上,结合复杂事件处理技术在实时处理方面的优势,开展基于CEP并行化处理的犯罪线索实时推荐的研究和系统的实现。通过对职务犯罪类型之间的关系进行分析,构建了面向职务犯罪的关键词语义树,节点所在的层数作为节点的权值。本文根据关键词的语义关系来计算匹配的关键词个数和总权值,推荐规则是总权值的大小进行推荐。在模式分析的基础上,给出了模式研究与设计过程中的一些关键步骤,并针对系统的核心模块进行详细的设计与实现。经过实验结果对比分析,该方法比传统方法的准确率要高。在海量数据和系统实时性要求比较高时,该系统具有一定的优势。
[Abstract]:Throughout history, crime has been accompanied by it from the date of the emergence of the country, existing in various historical stages and various social types, and has never disappeared. The birth and development of network technology provides a broader way for case investigation. Relying on the network to collect criminal clues will become a new method of case investigation. With the rapid development of Internet technology, the information data of the network has shown an exponential growth, so how to quickly find valuable clues from the massive data and recommend it to the relevant departments in real time is an urgent problem to be solved in case investigation. Complex event processing (Complex Event Processing,CEP) is the process of abstracting the external flow of data into events, and then detecting and responding to the pre-defined complex event patterns through the filtering and aggregation of events. Its biggest advantage is to process real-time data, which allows us to get the most useful information in the shortest possible time. The parallelization of complex event processing effectively solves the performance problems of centralized processing and shows great advantages in the face of massive data. On the basis of the collection of the original crime clues, the parallel complex event processing technology is introduced to provide the real-time and efficient guarantee for the relevant departments to recommend the crime clues. In this paper, through the research of complex event processing parallelization technology, on the basis of the existing network crime clue collection system, combined with the advantages of complex event processing technology in real-time processing, The research and system implementation of real-time recommendation of crime clues based on CEP parallel processing are carried out. Based on the analysis of the relationship between the types of job-related crimes, a keyword semantic tree for job-related crimes is constructed, and the number of layers in which the nodes are located is taken as the weight of the nodes. In this paper, the number and total weight of matching keywords are calculated according to the semantic relationship of keywords, and the recommendation rule is the size of total weights to recommend. On the basis of pattern analysis, some key steps in the process of pattern research and design are given, and the core modules of the system are designed and implemented in detail. The experimental results show that the accuracy of this method is higher than that of the traditional method. When the real-time requirements of massive data and system are relatively high, the system has certain advantages.
【学位授予单位】:华北电力大学(北京)
【学位级别】:硕士
【学位授予年份】:2017
【分类号】:D917
本文编号:2480453
[Abstract]:Throughout history, crime has been accompanied by it from the date of the emergence of the country, existing in various historical stages and various social types, and has never disappeared. The birth and development of network technology provides a broader way for case investigation. Relying on the network to collect criminal clues will become a new method of case investigation. With the rapid development of Internet technology, the information data of the network has shown an exponential growth, so how to quickly find valuable clues from the massive data and recommend it to the relevant departments in real time is an urgent problem to be solved in case investigation. Complex event processing (Complex Event Processing,CEP) is the process of abstracting the external flow of data into events, and then detecting and responding to the pre-defined complex event patterns through the filtering and aggregation of events. Its biggest advantage is to process real-time data, which allows us to get the most useful information in the shortest possible time. The parallelization of complex event processing effectively solves the performance problems of centralized processing and shows great advantages in the face of massive data. On the basis of the collection of the original crime clues, the parallel complex event processing technology is introduced to provide the real-time and efficient guarantee for the relevant departments to recommend the crime clues. In this paper, through the research of complex event processing parallelization technology, on the basis of the existing network crime clue collection system, combined with the advantages of complex event processing technology in real-time processing, The research and system implementation of real-time recommendation of crime clues based on CEP parallel processing are carried out. Based on the analysis of the relationship between the types of job-related crimes, a keyword semantic tree for job-related crimes is constructed, and the number of layers in which the nodes are located is taken as the weight of the nodes. In this paper, the number and total weight of matching keywords are calculated according to the semantic relationship of keywords, and the recommendation rule is the size of total weights to recommend. On the basis of pattern analysis, some key steps in the process of pattern research and design are given, and the core modules of the system are designed and implemented in detail. The experimental results show that the accuracy of this method is higher than that of the traditional method. When the real-time requirements of massive data and system are relatively high, the system has certain advantages.
【学位授予单位】:华北电力大学(北京)
【学位级别】:硕士
【学位授予年份】:2017
【分类号】:D917
【引证文献】
相关硕士学位论文 前1条
1 万傲青;基于大众点评网数据的武汉中心城区餐饮消费空间分布研究[D];湖北大学;2018年
,本文编号:2480453
本文链接:https://www.wllwen.com/falvlunwen/fanzuizhian/2480453.html