Knowledge-intensive workers, such as academic researchers, medical professionals or patent engineers, have a demanding need of searching information relevant to their work. Content-based recommender system (CBRS) makes recommendation by analyzing similarity of textual contents between documents and users’ preferences. Although content-based filtering has been one of the promising approaches to document recommendations, it encounters the over-specialization problem. CBRS tends to recommend documents that are similar to what have been in user’s preference profile. Rationally, citations in an article represent the intellectual/affective balance of the individual interpretation in time and domain understanding. A cited article shall be associated with and may reflect the subject domain of its citing articles. Our study addresses the over-specialization problem to support the information needs of researchers. We propose a Reference Topic-based Document Recommendation (RTDR) technique, which exploits the citation information of a focal user’s preferred documents and thereby recommends documents that are relevant to the subject domain of his or her preference. Our primary evaluation results suggest the outperformance of the proposed RTDR to the benchmarks.
|出版狀態||已出版 - 2017|
|事件||21st Pacific Asia Conference on Information Systems: Societal Transformation Through IS/IT, PACIS 2017 - Langkawi, Malaysia|
持續時間: 16 7月 2017 → 20 7月 2017
|???event.eventtypes.event.conference???||21st Pacific Asia Conference on Information Systems: Societal Transformation Through IS/IT, PACIS 2017|
|期間||16/07/17 → 20/07/17|