Learning to rank for question routing in community question answering

Abstract

This paper focuses on the problem of Question Routing (QR) in Community Question Answering (CQA), which aims to route newly posted questions to the potential answerers who are most likely to answer them. Traditional methods to solve this problem only consider the text similarity features between the newly posted question and the user profile, while ignoring the important statistical features, including the question-specific statistical feature and the user-specific statistical features. Moreover, traditional methods are based on unsupervised learning, which is not easy to introduce the rich features into them. This paper proposes a general framework based on the learning to rank concepts for QR. Training sets consist of triples (q, asker, answerers) are first collected. Then, by introducing the intrinsic relationships between the asker and the answerers in each CQA session to capture the intrinsic labels/orders of the users about their expertise degree of the question q, two different methods, including the SVM-based and RankingSVM-based methods, are presented to learn the models with different example creation processes from the training set. Finally, the potential answerers are ranked using the trained models. Extensive experiments conducted on a real world CQA dataset from Stack Overflow show that our proposed two methods can both outperform the traditional query likelihood language model (QLLM) as well as the state-of-the-art Latent Dirichlet Allocation based model (LDA). Specifically, the RankingSVMbased method achieves statistical significant improvements over the SVM-based method and has gained the best performance.

A Survey on Expert Recommendation in Community Question Answering

“…Zhou et al [66] also use SVM but they define both local and global features on questions, user history, and question-user relationship and additional consider KLdivergence as a new feature. Ji et al [67] additionally use text similarities as features to train SVM and one of its variant, RankingSVM.…”

“…Given the advantages of ranked recommendation results over unranked results, Ji et al [67] propose Rank-ingSVM, a ranking model based on SVM, for the expert recommendation. Burel et al [72] extract patterns from the question-selection behaviors of users in a Q&A community and then use Learning to Rank (LTR) models to identify the most relevant question to a user at any given time.…”

Personalized Question Routing via Heterogeneous Network Embedding

“…The answer collection is consequently accelerated since these users are able to immediately spot the questions of their expertise. Such task is also known as question routing and is previously addressed by feature engineering-based approaches (Zhou, Lyu, and King 2012;Ji and Wang 2013;Chang and Pal 2013). Features exploited include the statistics of users, the language modeling features of question content, and the relationships between users and questions.…”

“…Such ranking score measures the probability of the answerer providing the "accepted answer" to this question. Compared with previous frameworks (Zhou, Lyu, and King 2012;Zhao et al 2015;Ji and Wang 2013), our ranking function explicitly computes the ranking scores taking advantage of rich non-linear information of the three entities.…”

Survey Paper on a Pairwise Learning to Rank Model for Answer Selection in Community Question Answer (CQA) System

“…[9] Z. Ji and B. Wang(2013) has represented Learning to rank for question routing in community question answering [9]. The author proposes a general framework based on the learning to rank concepts for QR.…”