苏州科技学院学报(自然科学版)
囌州科技學院學報(自然科學版)
소주과기학원학보(자연과학판)
JOURNAL OF UNIVERSITY OF SCIENCE AND TECHNOLOGY OF SUZHOU(NATURAL SCIENCE EDITION)
2014年
3期
75-80
,共6页
推荐系统%协同过滤%数字图书馆%推荐算法
推薦繫統%協同過濾%數字圖書館%推薦算法
추천계통%협동과려%수자도서관%추천산법
recommender system%collaborative filtering%digital library%recommender algorithm
协同过滤算法应用于数字图书馆推荐系统有诸多优点,然而在系统启动初期存在最初评价问题和数据稀疏问题。针对这些情况,提出引入文献引证关系于数字图书馆协同过滤算法。文献引证关系能够较好的反映文献作者对参考文献、引证文献的兴趣评价。基于文献引证关系的数字图书馆协同过滤算法引入作者的文献评价信息,解决了协同过滤算法固有的问题。实验结果表明,该算法能有效地提高系统的推荐质量。
協同過濾算法應用于數字圖書館推薦繫統有諸多優點,然而在繫統啟動初期存在最初評價問題和數據稀疏問題。針對這些情況,提齣引入文獻引證關繫于數字圖書館協同過濾算法。文獻引證關繫能夠較好的反映文獻作者對參攷文獻、引證文獻的興趣評價。基于文獻引證關繫的數字圖書館協同過濾算法引入作者的文獻評價信息,解決瞭協同過濾算法固有的問題。實驗結果錶明,該算法能有效地提高繫統的推薦質量。
협동과려산법응용우수자도서관추천계통유제다우점,연이재계통계동초기존재최초평개문제화수거희소문제。침대저사정황,제출인입문헌인증관계우수자도서관협동과려산법。문헌인증관계능구교호적반영문헌작자대삼고문헌、인증문헌적흥취평개。기우문헌인증관계적수자도서관협동과려산법인입작자적문헌평개신식,해결료협동과려산법고유적문제。실험결과표명,해산법능유효지제고계통적추천질량。
Collaborative filtering algorithm in digital library recommender system has many advantages, but in the initial system startup, there were two major problems, including the initial evaluation and data sparsity. In view of this, we propose to introduce the literature citation relations to the digital library collaborative filtering algorithm. The relations can represent the literature evaluation of authors interested in reference of the literature and citation of the literature. Digital library collaborative filtering recommender system based on citation relations solves the inherent problems of collaborative filtering algorithm by introducing evaluations of the authors. The ex-perimental results show that the algorithm can effectively improve the quality of the recommender system.