Collaborative filtering recommendation algorithm based on multi-level hybrid similarity

doi:10.11772/j.issn.1001-9081.2017071718

Abstract

Abstract: In view of performance flaws in the case of sparse data and the lack of similarity measurement methods in traditional collaborative filtering recommendation algorithm, a collaborative filtering recommendation algorithm based on multi-level hybrid similarity was proposed to improve the recommendation accuracy. The algorithm is divided into three different levels. Firstly, the concept of fuzzy set was used to fuzzify the user rating and then to calculate the user's fuzzy preferences, and the adjusted cosine-based similarity of the user rating and the Jarccad similarity of the user rating were combined as the user rating similarity. Secondly, the use rating was classified to predict the degree of interest of the user to the item category so that the user's interest similarity was calculated. Thirdly, the user's characteristic similarity was predicted by the characteristic attributes between users. Then, the user's interest similarity and user's characteristic similarity were dynamically integrated by the number of user ratings. Finally, the similarities of three levels were fused as the result of user similarity. The experimental results show that the improved hybrid algorithm has a decrease of 5% in Mean Absolute Error (MAE) compared to the adjusted cosine-based similarity algorithm when the number of neighbors is small. Compared with the improved MKJCF (Modified K-pow Jaccard similarity Cooperative Filtering) algorithm, the improved hybrid algorithm has a slight advantage, and the MAE fell by an average of about 1% with the increase of neighbor number. The proposed algorithm uses a multi-level recommendation strategy to improve the user's recommendation accuracy, effectively alleviates the sparseness of data and the impact of single measurement method.

Key words: collaborative filtering, data sparseness, fuzzy set, rating similarity, interest similarity, characteristic similarity

摘要： 针对传统协同过滤推荐算法在数据稀疏的情况下存在的性能缺陷和相似性度量方法的不足，为了提高推荐精度，改进原算法得到了一种基于多层次混合相似度的协同过滤推荐算法。该算法主要分为三个不同的层次：首先采用模糊集的概念将用户评分模糊化，计算用户的模糊偏好，并结合用户评分的修正余弦相似度和用户评分的Jarccad相似度总体作为用户评分相似度；再对用户评分进行分类来预测用户对项目类别的兴趣程度，从而计算出用户兴趣相似度；然后利用用户的特征属性来预测用户之间的特征相似度；其次根据用户评分数量来动态地融合用户兴趣相似度及用户特征相似度；最后融合三个层次的相似度作为用户混合相似度的结果。利用MovieLens公用数据集对改进前后的算法进行对比实验，结果表明：当在邻居集合数量较少时，改进的混合算法相对修正余弦相似度算法的平均绝对偏差（MAE）下降了5%左右；较改进的修正的Jaccard相似性系数的协同过滤（MKJCF）算法也存在略微的优势，随着邻居集合数的增加MAE也平均下降了1%左右。该算法采用多层次的推荐策略提高了用户的推荐精度，有效地缓解了数据稀疏性问题和单一度量方法的影响。

关键词: 协同过滤, 数据稀疏性, 模糊集, 评分相似度, 兴趣相似度, 特征相似度

CLC Number:

TP391.4

YUAN Zhengwu, CHEN Ran. Collaborative filtering recommendation algorithm based on multi-level hybrid similarity[J]. Journal of Computer Applications, 2018, 38(3): 633-638.

袁正午, 陈然. 基于多层次混合相似度的协同过滤推荐算法[J]. 计算机应用, 2018, 38(3): 633-638.

References

[1] 王娜,任婷.移动社交网站中的信息过载与个性化推荐机制研究[J]. 情报杂志,2015,34(8):190-194.(WANG N, REN T. Information overload and personalized recommendations mechanism in the mobile social networking sites[J]. Journal of Intelligence, 2015,34(8):190-194.)
[2] SAWAR B, KARYPIS G, KONSTAN J, et al. Item-based collaborative filtering recommendation algorithms[C]//WWW'01:Proceedings of the 10th International Conference on World Wide Web. New York:ACM, 2001:285-295.
[3] 孙小华.协同过滤系统的稀疏性与冷启动问题研究[D].杭州:浙江大学,2005.(SUN X H. Research on sparseness and cold start of collaborative filtering system[D]. Hangzhou:Zhejiang University, 2005.)
[4] 冷亚军,陆青,梁昌勇.协同过滤推荐技术综述[J].模式识别与人工智能,2014,27(8):720-734.(LENG Y J, LU Q, LIANG C Y. Survey of recommendation based on collaborative filtering[J]. Pattern Recognition and Artificial Intelligence, 2014, 27(8):720-734.)
[5] 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.(HUANG C G, YIN J, WANG J, et al. Uncertain neighbors' collaborative filtering recommendation algorithm[J]. Chinese Journal of Computers, 2010, 33(8):1369-1377.)
[6] 时念云,葛晓伟,马力.基于用户人口统计特征与信任机制的协同推荐[J]. 计算机工程, 2016,42(6):180-184.(SHI N Y, GE X W, MA L. Collaborative recommendation based on user demographics and trust mechanism[J]. Computer Engineering, 2016, 42(6):180-184.)
[7] YU K, XU X, ESTER M, et al. Feature weighting and instance selection for collaborative filtering:an information-theoretic approach[J]. Knowledge and Information Systems, 2003, 5(2):201-224.
[8] 任看看,钱雪忠.协同过滤算法中的用户相似性度量方法研究[J].计算机工程,2015,41(8):18-22.(REN K K, QIAN X Z. Research on user similarity measure method in collaborative filtering algorithm[J]. Computer Engineering, 2015, 41(8):18-22.)
[9] AL-SHAMRI M Y H, AL-ASHWAL N H. Fuzzy-weighted similarity measures for memory-based collaborative recommender systems[J]. Journal of Intelligent Learning Systems and Applications, 2014, 6(1):1-10.
[10] 陈雪,黄智力,罗键.基于相对相似度关系的三角模糊数型不确定多属性决策法[J].控制与决策,2016,31(12):2232-2240.(CHEN X, HUANG Z L, LUO J. Approach for triangular fuzzy number-based uncertain multi-attribute decision making based on relative similarity degree relation[J]. Control and Decision, 2016, 31(12):2232-2240.)
[11] 胡伟健,滕飞,李灵芳.适应用户兴趣变化的改进型协同过滤算法[J]. 计算机应用,2016,36(8):2087-2091.(HU W J, TENG F, LI L F. Improved adaptive collaborative filtering algorithm to change of user interest[J]. Journal of Computer Applications, 2016, 36(8):2087-2091.)
[12] 贾冬艳,张付志.基于双重邻居选取策略的协同过滤推荐算法[J].计算机研究与发展,2013,50(5):1076-1084.(JIA D Y, ZHANG F Z. A collaborative filtering recommendation algorithm based on double neighbor choosing strategy[J]. Journal of Computer Research and Development, 2013, 50(5):1076-1084.)
[13] 王明佳,韩景倜.基于用户对项目属性偏好的协同过滤算法[J].计算机工程与应用,2017,53(6):106-110.(WANG M J, HAN J T. Collaborative filtering algorithm based on item attribute preference[J]. Computer Engineering and Applications, 2017, 53(6):106-110.)
[14] POLATIDIS N, GEORGIADIS C K. A multi-level collaborative filtering method that improves recommendations[J]. Expert Systems with Applications, 2016,48:100-110.
[15] 张南,林晓勇,史晟辉.基于改进型启发式相似度模型的协同过滤推荐方法[J].计算机应用,2016,36(8):2246-2251.(ZHANG N, LIN X Y, SHI S H. Collaborative filtering recommendation method based on improved heuristic similarity model[J]. Journal of Computer Applications, 2016, 36(8):2246-2251.)