ELS: An Efficient Entity Linking System

Abstract

WISE 2013 Challenge T1 is to label entities within plain texts based on a given set of entities. The set of entities used for labeling is extracted from the Wikilinks dataset. For the given test corpus (in English) and a list of entities, the challenge will automatically label the mentioned entities or concrete concepts within the given test corpus. In this report, we present ELS, which is an efficient entity linking system that can detect the entities effectively. Firstly,we convert each file into a list of tokens. In order to make sure the labeled concepts are concrete,the windows model is designed to extend the tokens. As the same entity may be linked by different URLs,we select the optimal linking via edit distance. Finally, we conduct extensive experiments to verify the efficiency of our proposed ELS.