Paper: A Co-occurrence Graph-based Approach for Personal Name Alias Extraction from Anchor Texts

ACL ID I08-2123
Title A Co-occurrence Graph-based Approach for Personal Name Alias Extraction from Anchor Texts
Venue International Joint Conference on Natural Language Processing
Session Main Conference
Year 2008
Authors

A person may have multiple name aliases on the Web. Identifying aliases of a name is important for various tasks such as in- formation retrieval, sentiment analysis and name disambiguation. We introduce the no- tion of a word co-occurrence graph to rep- resent the mutual relations between words that appear in anchor texts. Words in an- chor texts are represented as nodes in the co-occurrence graph and an edge is formed between nodes which link to the same url. For a given personal name, its neighboring nodes in the graph are considered as can- didates of its aliases. We formalize alias identification as a problem of ranking nodes in this graph with respect to a given name. We integrate various ranking scores through support vector machines to leverage a robust ranking function and use it t...