Abstract:

Person′s names are so ambiguous that the results of searching for a person′s name are usually a mixture of pages about namesakes. Person′s name disambiguation is a course of distinguishing different person′s entities with the same name. The method of person′s name disambiguation based on the relevant community was proposed and the modified Espresso algorithm was used to find relevant community for each Web page. The enlarged name sets were applied in the two-stage person′s name disambiguation algorithm, and then the algorithm was tested it on the WePS-2 test dataset. The experimental results show the effectiveness of our method.

Key words: social network; community; person′s name disambiguation; Web people search; clustering