基于中文Web社会网络的提取、测量与分析

从中文Web中提取人名实体,设计了3种人物间关系定义,并以此为基础构建一个人物间关系网络。通过计算一些社会网络研究中常用的属性指标,验证了社会网络中的一些常见现象,例如“小世界现象”。对社会网络的分析工作包括两个方面:一方面,使用主成份分析来帮助划分人物关系网络,并将划分结果与对应的人物属性相比较,结果表明人物职业较之籍贯和性别等属性更符合来自Web社会网络中的结构划分;另一方面,根据人名在Web中热点程度排序,并按照顺序依次选择3组不同大小的人物列表分别构建社会网络,以此来观察随着人物的增加该社会网络的演化现象,并发现构建的人物关系网络结构是一个单核心的网络。

Extraction,Measurement and Analysis of Social Network in Chinese Web

In this article,people's names are extracted from Chinese Web,and three measurements of their relationship are designed to build a relationship network.Several measures can validate some phenomenon such as "small world" phenomenon.Two analytical tasks should be included in this network:one is to cluster the names by applying PCA to the network and compare the result with a manual labeled person list.The result shows that occupation attributs better matches to the cluster result than other factors such as gender and birth place.The other is to construct a network according to the order of the three different sizes of name list in Chinese Web.It can be found,and it finds that the relationship network has only one core.