| 基于聚类的反恐情报异常数据分析方法研究 |
| |
| 引用本文： | 李勇男.基于聚类的反恐情报异常数据分析方法研究[J].现代情报,2019,39(10):32. |
| |
| 作者姓名： | 李勇男 |
| |
| 作者单位： | 中国人民公安大学, 北京 100038 |
| |
| 基金项目： | 国家自然科学基金项目"基于DNA自组装并行计算模型的二元域圆锥曲线密码编码机理研究"（项目编号：61702523）；北京市社会科学基金项目"大数据驱动的首都反恐情报决策机制研究"（项目编号：18GLC062）；教育部人文社会科学研究青年基金项目"基于数据挖掘的涉恐情报量化分析方法研究"（项目编号：17YJCZH098）。 |
| |
| 摘 要： | 目的/意义]通过异常检测可以在海量涉恐数据中发现异常信息，为反恐预警提供重要情报。方法/过程]首先利用聚类将基础数据分为不同的簇，识别出其中区别于大部分数据对象的异常人员，然后设计一种专门的相似度综合度量参数用于计算与恐怖分子最相似的人员。结果/结论]该方法为检测异常人员数据提供了一种可以参考的思路，用于从多种来源数据中快速找出涉恐敏感程度较高的人员，有望提高反恐情报分析的效率，实现精准打击重点涉恐人员和恐怖活动。
|
| 关 键 词： | 数据挖掘 异常检测 聚类分析 相似度 反恐情报 |
Research on Outlier Detection in the field of Counter Terrorism Intelligence Analysis based on Clustering |
| |
| Authors: | Li Yongnan |
| |
| Institution: | People's Public Security University of China, Beijing 100038, China |
| |
| Abstract: | Purpose/Significance]Outlier detection from mass fundamental data could provide important information about latent terrorists for early warning of counter-terrorism.Method/Process]Abnormal people that differed from most data objects must be identified by using clustering method to classify mass terror related data with composite properties.Additionally,abnormal people acted as the basic data source to find terror related people.These subjects who were excavated out had the larger degree of similarity with terrorists.Result/Conclusion]This method provided an idea to detect high sensitivity people from multiple data streams.It could accelerate the speed of handling counter-terrorism intelligence and provide reference for counter-terrorism disposition by means of swiftly finding the terror related people. |
| |
| Keywords: | data mining outlier detection clustering degree of similarity counter terrorism intelligence |
|
| 点击此处可从《现代情报》浏览原始摘要信息 |
| 点击此处可从《现代情报》下载免费的PDF全文 |
