Finding Experts in Community Question Answering System Using Trie String Matching Algorithm with Domain Knowledge

Abstract

In recent days, different community question-answering systems have been evolving as online forums for knowledge sharing among users across the world. However, finding experts on the questions is a key issue in the community question-answering system. To address this issue, a novel approach is proposed in this paper for finding experts from the community question-answering website using an exact string-matching algorithm along with domain knowledge. The proposed system is designed with three phases: i) User Profile Modeling, ii) Question Preprocessing, and iii) Expert Recommendation. Initially, the matrix factorization model is adopted for user profile modeling where the tags and the answerer’s information are represented in the form of a matrix. Then, the domain-wise grouping of tags is done to minimize the search time of the tags. The question preprocessing is done using a keyword extraction algorithm to extract the keywords. Finally, the expert recommendation is accomplished using the trie string matching algorithm and key-value mapping process. For doing experiments, a stack overflow community question-answering website is utilized in this work. The performance of the system is measured and the results section proved that the system achieved 91.2% accuracy.