Author Name Disambiguation for Citations on the Deep Web

Rui Zhang, Derong Shen, Yue Kou, Tiezheng Nie

Abstract: Name ambiguity is a critical problem in many applications, in particular in the online bibliographic digital libraries. Although several clustering-based methods have been proposed, the problem still presents to be a big challenge for both data integration and cleaning process. In this paper, we present a complementary study to the author name disambiguation from another point of view. We focus on the common names, especially non-canonical ones. We propose an approach of automatic access to authors’ personal information over Deep Web, and compute the similarity of every two citations according to the following features: co-author name, author’s affiliation, e-mail address and title. Then we employ Affinity Propagation clustering algorithm to attributing the resembling citations to the proper authors. We conducted experiments based on five data sources: DBLP, CiteSeer, IEEE, ACM and Springer LINK. Experiments results show that significant improvements can be obtained by using the proposed approach.