计算机科学 ›› 2014, Vol. 41 ›› Issue (1): 69-71.
蔡强,韩东梅,李海生,胡耀光,陈谊
CAI Qiang,HAN Dong-mei,LI Hai-sheng,HU Yao-guang and CHEN Yi
摘要: 传统的协同过滤算法以用户评分体现用户兴趣偏好及资源相似度,忽视了用户、资源自身的特征,并且对稀疏数据和新资源的推荐质量明显下降。在Web2.0时代下,标签可被用户依个人偏好进行自由资源标注。因此,提出了基于标签和协同过滤的推荐算法。其基本思想是将标签作为体现用户兴趣偏好和资源特征的信息,依据用户、标签及资源的多维关系生成用户及资源的标签特征向量,并计算用户对资源的偏好程度和资源相似度,然后基于用户的历史行为预测用户对其他资源的偏好值,最后依据预测偏好值排序产生Top-N推荐结果。通过与传统的协同过滤算法的比较,验证了本算法能有效缓解数据的稀疏性,解决推荐的冷启动问题,提升推荐的准确性,获得更好的推荐效果。
[1] Kohi A,Ebrahimi S J,Jalali M.Improving the accuracy and efficiency of tag recommendation system by applying hybrid me-thods.comper[C]∥20111st International eConference on Computer and Knowledge Engineering.Mashhad,Iran,2011:242-248 [2] 张斌,张引,高克宁,等.融合关系与内容分析的社会标签推荐[J].软件学报,2012,23(3):476-488 [3] 李聪,梁昌勇,马丽.基于协同过滤与划分聚类的改进推荐算法[J].计算机研究与发展,2008,45(9):1552-1538 [4] 黄创光,印鉴,汪静,等.不确定近邻的协同过滤推荐算法[J].计算机学报,2010,33(8):1369-1377 [5] Koren Y,Bell R,Volinsky C.Matrix Factorization techniques for recommender systems[J].IEEE Computer Society,2009,42(8):30-37 [6] Jin Jian,Chen Qun.A Trust-based top-k recommender systemusing social tagging network[C]∥20129th International Conference on Fuzzy Systems and Knowledge Discovery.China,2012:1270-1274 [7] Nanopoulos A,Rafailidis D,Symeonidis P,et al.MusicBox:personalized music recommendation based on cubic analysis of social tags[J].IEEE Transaction on Audio,Speech,and Language Processing,2010,18(2):407-412 [8] Rau Jer-wei,Huang Jen-wei,Yung Sheng.Improving the qualityof tags using state transition on progressive image search and recommendation system[C]∥2012IEEE International Conference on Systems,Man,and Cybernetics.Seoul,2012:3233-3238 [9] Song Yang,Lu Zhang.Automatic tag recommendation algo-rithms for social recommender systems[J].ACM Transaction on the Web.2011,5(1):1-31 [10] Xia Xiu-feng,Zhang Shu,Li Xiao-ming.A personalized recommendation model based on social tags[C]∥International Workshop on Database Technology and Applications.Wuhan,2010:1-5 [11] 韦素云,业宁,朱健,等.基于资源聚类的全局最近邻的协同过滤算法[J].计算机科学,2012,39(12):149-152 [12] Hao Fei,Zhong Sheng-tong.Tag recommendation based on user interest lattice matching[C]∥IEEE International Conference on Computer Science and Information Technology.Daejeon,2010:276-280 [13] Olvera E P,Godoy D.Valuating term weighting schemes forcontent-based tag recommendation in social tagging systems[J].IEEE Latin America Transaction,2012,10(4):1973-1980 |
No related articles found! |
|