Person Name Disambiguation based on Topic Model

In this paper we describe our participation in the SIGHAN 2010 Task3 (Person Name Disambiguation) and detail our approaches. Person Name Disambiguation is typically viewed as an unsupervised clustering problem where the aim is to partition a name’s contexts into different clusters, each representing a real world people. The key point of Clustering is the similarity measure of context, which depends upon the features selection and representation. Two clustering algorithms, HAC and DBSCAN, are investigated in our system. The experiments show that the topic features learned by LDA outperforms token features and more robust.