文章基本信息

标题：An Effective Academic Research Papers Recommendation for Non-profiled Users
本地全文：下载
作者：Damien Hanyurwimfura ; Liao Bo ; Vincent Havyarimana 等
期刊名称：International Journal of Hybrid Information Technology
印刷版ISSN：1738-9968
出版年度：2015
卷号：8
期号：3
页码：255-272
DOI：10.14257/ijhit.2015.8.3.23
出版社：SERSC
摘要：With the tremendous amount of research publications online, finding relevant ones for a particular research topic can be an overwhelming task. As a solution, papers recommender systems have been proposed to help researchers find their interested papers or related papers to their fields. Most of existing papers recommendation approaches are based on paper collections, citations and user profile which is not always available (not all users are registered with their profiles). The existing approaches assume that users have already published papers and registered in their systems. Consequently, this neglects new researcher without published papers or profiles. In this paper, we propose an academic researcher papers recommendation approach that is based on the paper's topics and paper's main ideas. The approach requires as input only a single research paper and extracts its topics as short queries and main ideas' sentences as long queries which are then submitted to existing online repositories that contains research papers to retrieve similar papers for recommendation. Four query extraction and one paper recommendation methods are proposed. Conducted experiments show that the proposed method presents good improvement.
关键词：academic paper recommendation; topics extraction; paper relationship; ; multi words topics; cosine similarity