Treatment of Social Media in Person Name Disambiguation in the Web

In this work, we present two heuristics to treat web pages from social networks for person name disambiguation in the Web. This problem consists in clustering the results provided by a search engine when the query is a person name according to the individual they refer to. Although these web pages could negatively affect when grouping the results, most of the systems in the state-of-the-art do not take into account their role in this problem. We have evaluated our heuristics with two collections that contain this kind of web pages. We have used an extension of an algorithm of the state of the art to cluster the web pages. Both heuristics get improvements when there is a high number of social web pages, and the proposed algorithm is more independent with respect to the ambiguity degree of person names than other ones in the state of the art.