Citation Context

...and to leverage computational techniques to improve the user experience. For example, researchers have investigated the use of machine learning techniques to predict askers' satisfaction with answers =-=[15]-=-, build models of question and answer quality based on features derived from text and user relationships [2], predict the occurrence of “best answers” [1], and compute metrics of expected user quality...

Citation Context

...for identifying users with high expertise [10, 17, 25, 26]. The second is the perspective of information retrieval where a question is viewed as a “query” and answers could be thought of as “results” =-=[15, 8, 16, 2, 9]-=-. One goal of this line of work is to take a question with multiple answers and extract the answer of best quality or the answer that is most related to a particular search query. This can be viewed a...

Citation Context

...ent [2, 4] and improve content quality [8]. Yet other studies try to predict certain properties of Q&A forums (e.g., the likelihood an answer will be chosen as a best answer [1] or asker satisfaction =-=[13]-=-). Online question and answer forums fit into the larger realm of peer production systems [3], which is a term used to refer to decentralized system of users that contribute to a system to achieve a g...

Citation Context

...rs, and proposed a language modeling-based retrieval model for processing these features in order to predict the quality of answers collected from a specific CQA service. Agichtein and his colleagues =-=[1, 23]-=- made great efforts on finding powerful features including structural, textual, and community features, and proposed a general classification framework for combining these heterogeneous features. Mean...

Citation Context

...eval. Similarly, Lam et al. [12] also focus on developing an automatic text categorization approach. Finally, we note that other works exists on CQA services that consider category information, e.g., =-=[1, 13]-=-. However, these focus on other aspects of CQA, not on question retrieval. 6. CONCLUSIONS Question retrieval is an essential component in Community Question Answer (CQA) services. In this paper, we pr...

Citation Context

...mportant to estimate user expertise score. Besides the question and answering community, user expertise score estimation is also studied in other social networks. Campbell et al. [5] used HITS to compute user expertise score over the user network of e-mail communications. Zhou et al. [24] followed the PageRank paradigm to propose an approach for co-ranking authors and their publications in a heterogeneous network. Research on user expertise score estimation can benefit a lot of applications. User expertise and other user related information are widely used for evaluating answer quality in CQA [1, 12, 17] and in online forums [7]. Suryanto et al. [22] incorporated user expertise score into question and answer search; they showed that user expertise score can improve question and answer search. Li et al. [16] proposed a framework to route questions to right answerers who are experts and available to answer questions. Horowitz et al. [11] developed a social search engine which routes questions to askers’ extended social network, including Facebook and Google Contacts, rather than the question and answering community. In this paper, our key idea is to propose a competition-based method that explo...

Citation Context

...uitable baseline. The closest previously addressed problem, at least in spirit, is asker satisfaction - predicting whether the original question asker (on the CQA site) was satisfied with the answers =-=[19]-=-. Intuitively, searcher satisfaction should be related to asker satisfaction. However, as we show later, asker satisfaction and searcher satisfaction appear to be very weakly related, if at all. Hence...

Citation Context

...s who asked at least 10 questions in that topic. Since our goal is to distinguish high quality questions from low quality ones, we follow the common binary classification setting in the previous work =-=[19, 15, 1]-=-. ] , Algorithm 1 MRLP-ST Input: user asking expertise vector U k 0 , question quality vector Q k 0, E, transition matrixes M and N, weighting coefficients α and β, some manual labels of U k 0 and/or ...

Citation Context

...y answers in the available CQA portals varies greatly. In many cases, the questions posted by askers go un-answered, or are answered poorly, never obtaining a satisfactory answer. In our recent work (=-=Liu et al., 2008-=-) we have introduced a general model for predicting asker satisfaction in community question answering. We found that previous asker history is a significant factor that correlates with satisfaction. ...

Citation Context

...1/03. high quality answers [8]. However, they can also have high latency even when the number of registered users is large [11]. Research on predicting answer quality and asker satisfaction (e.g., [8]=-=[18]-=-[21]) has typically been used a posteriori (e.g., to rate unrated answers) rather than predict outcomes as questions progress. Real-time Q&A services such as Aardvark [10], Twitter Answers (ask.mosio....