Article Preview
TopIntroduction
In the era of knowledge explosion, knowledge sharing services come up speedily. As a typical example, collaborative tagging systems (CTSs), such as CiteULike (http://www.flickr.com/), have already obtained the popularization use due to lower barrier and easier usability, considerable people join community activities. However, exploded knowledge raises many issues. Primarily, seekers have to spend plenty of time in searching the information that they are interested and the highly correlated experts (Sun et al., 2008). For example, for a search query ‘food’ on the web, there are thousands of options. Secondly, users often have different interests, levels of expertise or vocabulary bias, which can lead to tag synonyms, ambiguity or polysemy (Zhang et al., 2011b) due to inappropriate tags assigned to resources. Even worse, user may label tags arbitrarily for resources, resulting in noisy tags, and noisy tags cannot sufficiently describe the resource (Zhao et al., 2019). Eventually, these shortcomings weak the role of tags in information organization, sharing, retrieval and discovery. How to match the retrieve information with users’ expertise? To address these challenges, it is a novel solution to use expert finding and recommendation in CTSs.
Expert finding systems have been used in different objects, and many researchers have paid attention to this research field (Yang et al., 2007; Balog et al., 2009; Wang et al., 2013; Sun et al., 2015; Silva et al., 2013; Fang et al., 2008; Yuan et al., 2019). Recent decades, the common methods of the state-of-the-art techniques of EDAR are content analysis methods (Balog et al., 2009; Balog et al., 2006; Deng et al., 2008) and network analysis methods (Brin & Page,1998; Kleinberg et al., 1999; Bozzon et al. 2013, Zhou et al., 2007). For social networks, there are also similar applications of state-of-the-art techniques of EDAR, such as question answering communities and academia communicates. However, for CTSs such as Del.icio.us, social connections among users are hard to identify (Li et al., 2008). Therefore, it is an urgent research problem for expert finding and recommendation in CTSs with users' expertise? To address these challenges, it is a novel solution to use expert finding and recommendation in CTSs.