A social recommender mechanism for improving knowledge sharing in online forums

https://doi.org/10.1016/j.ipm.2011.10.004Get rights and content

Abstract

Nowadays, online forums have become a useful tool for knowledge management in Web-based technology. This study proposes a social recommender system which generates discussion thread and expert recommendations based on semantic similarity, profession and reliability, social intimacy and popularity, and social network-based Markov Chain (SNMC) models for knowledge sharing in online forum communities. The advantage of the proposed mechanism is its relatively comprehensive consideration of the aspects of knowledge sharing. Accordingly, results of our experiments show that with the support of the proposed recommendation mechanism, requesters in forums can easily find similar discussion threads to avoid spamming the same discussion. In addition, if the requesters cannot find qualified discussion threads, this mechanism provides a relatively efficient and active way to find the appropriate experts.

Highlights

► We build a social recommender system for knowledge sharing in online forums. ► The system finds relevant threads and experts efficiently to avoid spamming. ► The system integrates similarity, expertise, relation, and SNMC analysis modules. ► The system has better precision and recall rates than other benchmark methods.

Introduction

People exchange information and goods with one another in their daily life. Useful information becomes knowledge through face-to-face discussion. Document-recommendation based on locally stored files is proposed to aid users in efficiently discovering relevant information (Liu and Lai, 2009, Liu et al., 2008). With the vigorous development in Web-based technology, however, a large amount of information and knowledge has been transferred to online platforms, and the Internet provides an environment for people to contribute their knowledge and acquire the knowledge of others. One of the Internet’s intended purposes is to encourage knowledge sharing so that valuable knowledge embedded in the network can be effectively explored. Some previous studies on knowledge management show that the success of knowledge transfer lies in neither communication systems nor documents, but in social relationships (Argote et al., 2003, Ashok, 2005, Tobias and Finke, 2006, Yang and Chen, 2008). Knowledge sharing in some virtual communities, however, has not lived up to the expectation of Internet users in terms of acquiring and sharing valuable knowledge to meet their needs.

In knowledge-intensive online virtual communities such as technical discussion forums, knowledge sharing usually starts with knowledge inquiry in an online platform, which can take three forms: post-and-wait, know-what, and know-who (Wang, Chiang, Ho, Hsieh, & Huang, 2007). The “post-and-wait” approach describes anonymous queries where requesters, or searchers, post questions in a forum and wait for answers from volunteers. The “know-what” approach describes most situations in which huge volumes of documents or knowledge repositories are stored in a database ready to be retrieved by keyword matching or other more sophisticated yet similar techniques. The requester can only obtain what is stored in the database, and may have difficulty in evaluating the quality of the information acquired. The third approach, “know-who,” is to construct the connections between searchers and knowledge providers. The main requirement of this model is to possess knowledge as to who knows about the answer, and make it available to the searchers. The communities based on the “know-who” model will be more effective than the “post-and-wait” and “know-what” models, as they will facilitate knowledge transfer, and also enhance experiences and value exchange (Wang et al., 2007). Thus, the “know-who” model is a kind of expert find-and-discover model which can help accomplish knowledge sharing through social interaction.

Nowadays, online forums have become a useful tool for problem-solving (Stein & Maier, 1995), learning discussion (Jeng, Huang, Kuo, Chen, & Chu, 2005), and knowledge-building. They are established in several developed countries. In terms of countable posts, Japan is in the lead with over 2 million posts per day on their largest forum, 2channel.1 China also has many millions of posts on forums such as Tianya Club.2 The United States does not have any one large forum, but instead several hundred thousand smaller forums, the largest of which are Gaia Online,3 IGN,4 and GameFAQs5 (Wikipedia, 2009). Moreover, online forums prevail among various organizations. On the IBM website,6 there are seventeen online forums, over 30,000 threads and 200,000 replies. There are 28 online forums in the newly-established Microsoft research7 communities. This reveals the importance of the online forum for knowledge sharing. The most important benefit of online forums is that individual users can receive tailored answers from peers after formulating problems in their own words, without using specific keywords to search online (Steehouder, 2002). Inevitably, online forums face the same problematic questions with regard to knowledge sharing: “Who is the right expert to ask? When can I get the answer? How do I reduce the request time?” Without a practical means to ameliorate drawbacks and advance efficiency, the consequence is information-overload and time-wasting. A recommendation mechanism based on the “know-who” model is therefore proposed in this study.

There is great potential in online forums in the area of knowledge sharing, but Steehouder (2002) also argued that there are disadvantages in traditional forums, and we summarize these disadvantages as follows. First, if there is no participant interested in discussing it, a question will remain unanswered. Second, questions may remain in the discussion group for a long time before being answered. Third, the ways to reduce the waiting time for a question to be answered are limited. Considering these three disadvantages, this research proposes a social recommender system for more efficient knowledge sharing in online forum communities. In this study, we propose an innovative recommendation mechanism which identifies social relations and offers semantic analysis to construct a more comprehensive and personalized framework for each user in the online forum space, both for discussion threads and experts in the human-expert knowledge forum. Moreover, we utilize the concept of the Markov Chain model for social networks to find the most accessible and helpful experts if the user does not find satisfactory threads.

This proposed social recommender system provides the following advantages: First, the system can prevent the same question spamming the online forum. Second, the system can recommend potential experts to the requester if he/she cannot find the relevant answers. Third, the system can suggest the most willing experts to the requester. In certain knowledge forums, the help-seeker, called a requester, asks about a specific problem using natural language. The proposed model combines the aforementioned methods to find the relevant discussion threads. Nevertheless, if the search results do not meet the user’s needs, then the model recommends the most helpful experts and provides the shortest social path (including intermediate collaborators) for the requester. This model deals with the disadvantages of traditional online forum systems. The mechanism actively recommends the most helpful experts with willingness to solve the given problem. Accordingly, questions cannot be ignored and they are resolved in a short time via the shortest social path.

The remaining part of this paper is organized as follows. Section 2 introduces related works, followed by the whole framework in Section 3. Section 4 describes the experiments, data source, data statistics, and the results and evaluations of the experiments. Section 5 concludes and offers suggestions for further research.

Section snippets

Online forum

An online forum is a Web application for holding discussions and posting user-generated content in a specific domain, such as sports, recreation, techniques, and travel. People participating in an online forum may cultivate social bonds, and interest groups for a particular topic may form from the discussions. Forums contain a huge amount of valuable user-generated content on a variety of topics. It is highly desirable if the human knowledge contained in user-generated content in forums can be

The system framework

In this study, we propose a social recommender system which employs semantic analysis, expertise analysis, and relation analysis modules, as well as a modified Markov Chain model based on social networks to construct a more comprehensive and personalized framework for relevant discussion thread recommendation and expert-finding in the knowledge forum. Fig. 1 depicts the architecture of the proposed recommendation mechanism. There are four analysis modules developed to analyze the information

Experiments

In this section, we conduct empirical experiments using our proposed recommendation framework.

Conclusion

This paper proposes a social recommender system which includes semantic similarity, expertise, and social relation analyses for the support of knowledge sharing in online forum networks. The processes of recommendation include two phases: relevant discussion thread recommendation and helpful expert finding. In knowledge-centralized and highly diverse online forums, social intimacy and popularities improve recommendation effectiveness, as people who have the same interests and preferences easily

Acknowledgment

This research was supported by the National Science Council of Taiwan (Republic of China) under Grant NSC 97-2410-H-009-035-MY2.

References (47)

  • Balog, K., & Rijke, M. D. (2007). Determining expert profiles (with an application to expert finding). In Proceedings...
  • Balog, K., Azzopardi, L., Rijke, M. D., (2006). Formal models for expert finding in enterprise corpora. In Proceedings...
  • Blooma, M. J., Chua, A. Y. K., & Goh, D. H. L. (2008). A predictive framework for retrieving the best answer. In...
  • C.S. Campbell et al.

    Expertise identification using email communications

  • Chen, W., Zeng, Q., & Liu, W. (2006). A user reputation model for a user-interactive question answering system. In...
  • Cong, G., Wang, L., Lin, C. Y., Song, Y. I., & Sun, Y. (2008). Finding question–answer pairs from online forums. In...
  • R.M. Cooke et al.

    On the performance of social network and likelihood-based expert weighting schemes

    Reliability Engineering and System Safety

    (2008)
  • A. David et al.

    Actions, answers, and uncertainty: A decision-making perspective on Web-based question answering

    Information Processing & Management

    (2004)
  • Dom, B., Eiron, I., Cozzi, A., & Zhang, Y. (2003). Graph-based ranking algorithms for e-mail expertise analysis. In...
  • D.F. Ferraiolo et al.

    Proposed NIST standard for role-based access control

    ACM Transactions on Information and System Security

    (2001)
  • Fujimura, K., Inoue, T., & Sugisaki, M. (2005). The EigenRumor algorithm for ranking blogs. In Proceedings of the 2nd...
  • J. Golbeck

    Trust and nuanced profile similarity in online social networks

    Journal of Artificial Intelligence Research

    (2006)
  • W.W. Gunilla et al.

    Explaining knowledge sharing in organizations through the dimensions of social capital

    Journal of Information Science

    (2004)
  • Cited by (42)

    • Business process improvement and the knowledge flows that cross a private online social network: An insurance supply chain case

      2020, Information Processing and Management
      Citation Excerpt :

      Hung, Chen and Chung (2014) argue that, within ICT solutions, social networks are the most proficient knowledge sharing tools, since they are “web-based services that allow individuals to (1) construct a public or semi-public profile within a bounded system, (2) articulate a list of other users with whom they share a connection, and (3) view and traverse their list of connections and those made by others within the system” (Boyd & Ellison, 2008, p.211). Most studies that analyze knowledge sharing and its connection with social networks (Fang, Chi, Chen & Baron, 2015; Li, Liao & Lai, 2012; Massa, 2017; Xue, Bradley & Liang, 2011) adopt a subjective perspective; data involving knowledge flows is collected through a survey based on a questionnaire which only reflects recent events, and is the subject of cognitive biases (such as the anchoring effect, current moment bias, heuristic affect or omission bias) and circumstances. Although these approaches are in line with traditional knowledge management theories, which assume that knowledge sharing appears as the result of a rational decision, cognitive psychologists (Galinsky, Seiden, Kim & Medvec, 2002; Weingart et al., 1993) offer empirical evidence about the irrationality of human behavior in the context of knowledge sharing.

    • What motivates physicians to share free health information on online health platforms?

      2020, Information Processing and Management
      Citation Excerpt :

      Since the development of online technologies, many online communities and markets have been formed in the health care industry (Mein Goh, Gao & Agarwal, 2016). Such platforms provide a new way for patients to obtain health information and for physicians to provide the public with health and even medical information (Guo, Guo, Fang & Vogel, 2017; Li, Liao & Lai, 2012). Because the online approach can overcome geographic constraints and gives physicians greater access to information recipients (patients and their relatives), an increasing number have been using online platforms to share their professional knowledge (Wu & Po, 2016).

    • Experts and likely to be closed discussions in question and answer communities: An analytical overview

      2019, Computers in Human Behavior
      Citation Excerpt :

      However, as outlined in the conclusion, the demands on the dataset are very specific (e.g. relationships between users, direct messages), and they were unable to adequately illustrate the suitability of the tool as a quantitative evaluation method. Li, Liao, and Lai (2012) proposed a recommender system for the identification of experts, which is based on the profession of the user as well as perceived reliability, social intimacy, and popularity. Their study focused on examining social network-based Markov Chain models for knowledge sharing in online forum communities, and noted the need for a community-based process for the improvement of the recommendation process.

    • A context-aware researcher recommendation system for university-industry collaboration on R&D projects

      2017, Decision Support Systems
      Citation Excerpt :

      The new algorithm combines content-based expertise analysis and social network-based authority analysis to make suggestions. Li et al. combined semantic similarity and social relations for recommendations to improve knowledge sharing in online forum communities [16]. A research analytical framework (RAF) was proposed to recommend researchers for project selection [29].

    • Fostering scientists’ data sharing behaviors via data repositories, journal supplements, and personal communication methods

      2017, Information Processing and Management
      Citation Excerpt :

      Resource facilitating conditions were originally studied as external behavioral controls in the theory of planned behavior (Ajzen & Fishbein, 2005). Prior research reported that resource-facilitating conditions promote knowledge sharing and information sharing behaviors (Huang, Lin, & Chan, 2012; Li, Liao, & Lai, 2012). Data repositories can support data sharing in many STEM disciplines by providing data storage spaces and data management supports, such as electronic documents, digital archival collections, and data curation (Choudhury, 2008; Witt, 2008).

    View all citing articles on Scopus
    View full text