Extracting and Integrating Author Profiles from Research Publications

Abstract

In this paper we propose a mechanism for extraction and integration of research and work profiles of academic researchers based on a mix of a token-based and a character-based string similarity measure. The proposed technique uses Web mining techniques for extraction of publications metadata from digital libraries which is augmented from various online resources for the purposes of profile creation and integration. In particular we mine publications data from DBLP which is then populated with additional publications features obtained from the Web in a resource bound manner. On the basis of this populated publications data author profiles are extracted and integrated and presented them in terms of their important research attributes. These profiles present a clear picture of research output, research collaborations, research partners, frequency of publications with co-authors and shifting pattern of author affiliations, if any.