Domain expert ranking for finding domain authoritative users on community question answering sites

Abstract

Identification of experts is an important task in Community Question Answering (CQA) because many operations in CQA depend on the experts. In most of the CQA the expert identification techniques are very poor. The problem we identified is questions suggested by the cQa to the users who are identified as experts are not interested to answer by the identified experts. Also with current expert identification the experts are not identified in all domains. To tackle these problems, we propose a model Domain Expert Ranking (DER) which identifies the experts based on commonly available parameters in CQA. Experiments carried over on a question answering site Quora, shows that our model achieves considerable improvement over existing expert identification methods on multiple metrics.