Detecting Ambiguous Author Names in Crowdsourced Scholarly Data

The name ambiguity problem is a challenge in many areas, especially in the field of bibliographic digital libraries. For example, in services that use citation data to compute the impact of authors, ambiguous names lead to biased measures. The problem is amplified where names are collected from heterogeneous sources, including crowd sourced annotations. This is the case in the Scholaro meter system, which cross-correlates author names in user queries with those retrieved from bibliographic data. The uncontrolled nature of user-generated annotations is very valuable, but creates the need to detect ambiguous names. In this paper, we propose an approach to detect ambiguous names at query time, which makes it applicable in the context of a social computing application. We explore two kinds of heuristic features based on citations and crowd sourced topics. Our approach can detect ambiguous author names in crowd sourced scholarly data with an accuracy of 75%.