基于异构信息网络的假币犯罪信息分析系统  

Counterfeit Money Crimes Information Analysis System Based on Heterogeneous Information Networks

作　　者：陈贞虹 李慧[2] 许舒人[2] 叶丹[2] 

机构地区：[1]中国科学院大学,北京100049 [2]中国科学院软件研究所软件工程技术研究开发中心,北京100190

出　　处：《计算机系统应用》2016年第12期255-261,共7页Computer Systems & Applications

基　　金：国家自然科学基金(61379044);国家科技支撑课题(2015BAH18F02)

摘　　要：传统的犯罪查询的查询条件是文本信息,查询结果是有序的文档列表,这种方式无法展示结果之间的关系.基于异构信息网络以信息网络的形式重构假币犯罪信息数据,构建了假币犯罪信息网络,使用人名消歧的技术建立假币犯罪信息网络中嫌疑人之间的关系,并使用排序学习方法研究假币犯罪信息网络中的节点相关性问题,设计并实现了假币犯罪信息分析系统,通过以实体对象为查询项和网络图为查询结果的方式解决假币犯罪数据的查询问题.The query condition of traditional criminal query is a text, and query results are ordered documents lists. However, the way is not conducive to present the relationship of the results and help the police find clues of cases. Therefore, this paper reconstructs counterfeit crime data in the form of information network based on heterogeneous information network information, and constructs a counterfeit money crimes information network. We apply name disambiguation to build relationships of the suspects in counterfeit crime information network, and study relevancy problem of counterfeit money crimes information network using learning to rank methods. In this paper we design and implement the counterfeit money crimes information analysis system. To resolve the query problem of the criminal data, we use the entity as a query term and the network graph as the query results.

关 键 词：异构信息网络 元路径 犯罪数据挖掘 人名消歧 排序学习 

分 类 号：D918[政治法律—法学] TP393.09[自动化与计算机技术—计算机应用技术]