2015 | OriginalPaper | Buchkapitel
Cold-Start Expert Finding in Community Question Answering via Graph Regularization
verfasst von : Zhou Zhao, Furu Wei, Ming Zhou, Wilfred Ng
Erschienen in: Database Systems for Advanced Applications
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Expert finding for question answering is a challenging problem in Community-based Question Answering (CQA) systems such as Quora. The success of expert finding is important to many real applications such as question routing and identification of best answers. Currently, many approaches of expert findings rely heavily on the past question-answering activities of the users in order to build user models. However, the past question-answering activities of most users in real CQA systems are rather limited. We call the users who have only answered a small number of questions the cold-start users. Using the existing approaches, we find that it is difficult to address the cold-start issue in finding the experts.
In this paper, we formulate a new problem of cold-start expert finding in CQA systems. We first utilize the “following relations" between the users and topical interests to build the user-to-user graph in CQA systems. Next, we propose the
Graph Regularized Latent Model
(GRLM) to infer the expertise of users based on both past question-answering activities and an inferred user-to-user graph. We then devise an iterative variational method for inferring the GRLM model. We evaluate our method on a well-known question-answering system called Quora. Our empirical study shows encouraging results of the proposed algorithm in comparison to the state-of-the-art expert finding algorithms.