Unsupervised Web Name Disambiguation Using Semantic Similarity and Single-Pass Clustering

摘要

In this paper, we propose a method for name disambiguation. For a given set of names and documents we cluster the documents and map each cluster to the appropriate name. The proposed method incorporates an unsupervised metric for semantic similarity computation and a computationally low-cost clustering algorithm. We experimented with the data used in Web People Search Task of SemEval-2007, in which 16 different teams were participated. The proposed system has an equal performance compared to the officially best system.