Unsupervised Learning of Social Networks from a Multiple-Source News Corpus

Social Networks provide an intuitive picture of inferred relationships between people and organizations which allows different analyst tasks to be performed. In this paper we describe an unsupervised algorithm for learning of social networks from different news sources. The algorithm performs automatic paraphrase learning and multiple-source relation extraction to build a Social Network. We put forward a novel syntactic pattern matching algorithm which facilitates the scalability of our approach.

TANEV Hristo; 

2007-10-23

Incoma LTD

JRC37713

http://www-lipn.univ-paris13.fr/~poibeau/mmies/index.html,    https://publications.jrc.ec.europa.eu/repository/handle/JRC37713,