一种基于用户和物品相似度的融合协同过滤推荐算法
|
摘要
在协同过滤算法中,准确地找到相似用户是关键。稀疏数据问题常导致推荐性能低下,且推荐结果大都大量存在热门物品。文章基于物品(User-IIF)和用户(Item-IUF)的热门惩罚后的相似度计算方法,定义稀疏度的计算方式,并进行稀疏度加权来改善稀疏性问题,分别提出了两种融合后的协同过滤算法。改进后的两种融合算法更有效地利用了用户信息,提高了算法对降低热门物品干扰的效果,增加了推荐系统发现长尾物品的能力。实验采用MovieLens公开数据集,进行多次实验,实验结果表明改进后的两种融合算法在推荐结果的准确率和覆盖率上有显著提升,表明融合后的推荐算法在寻找相似用户或物品时表现得更高效。
In collaborative filtering algorithms, In the collaborative filtering algorithm, accurately finding similar users is the key.Sparse data problems often result in poor recommendation performance.In addition, most of the current recommendation systems ignore the recommendation for long tail items, and most of the recommended results are hot items. In this paper, we use the calculation method of similarity,the object-based(User-IIF) and user-based(Item-IUF),and define how sparsity is calculated, weighting sparsity to deal with data sparsity problem, and finally propose two fusion collaborative filtering algorithms. The improved two fusion algorithms can make more efficient use of users' information,and improve the effect of the algorithm on reducing the interference of popular items, increasing the ability of the recommendation system to find long tail items. In this paper, the experiment used the MovieLens public data set and several experiments have been carried out. The experimental results show that the improved two fusion algorithms have significantly improved the accuracy and coverage of the recommended results, indicating that the proposed recommendation algorithm is more efficient when searching for similar users or items.
In collaborative filtering algorithms, In the collaborative filtering algorithm, accurately finding similar users is the key.Sparse data problems often result in poor recommendation performance.In addition, most of the current recommendation systems ignore the recommendation for long tail items, and most of the recommended results are hot items. In this paper, we use the calculation method of similarity,the object-based(User-IIF) and user-based(Item-IUF),and define how sparsity is calculated, weighting sparsity to deal with data sparsity problem, and finally propose two fusion collaborative filtering algorithms. The improved two fusion algorithms can make more efficient use of users' information,and improve the effect of the algorithm on reducing the interference of popular items, increasing the ability of the recommendation system to find long tail items. In this paper, the experiment used the MovieLens public data set and several experiments have been carried out. The experimental results show that the improved two fusion algorithms have significantly improved the accuracy and coverage of the recommended results, indicating that the proposed recommendation algorithm is more efficient when searching for similar users or items.
引文
[1]黄立威,江碧涛,吕守业,等.基于深度学习的推荐系统研究综述[J].计算机学报,2018(41):1-30.
[2]荣辉桂,火生旭,胡春华,等.基于用户相似度的协同过滤推荐算法[J].通信学报,2014,35(2):16-24.
[3]Goldberg D,Oki B M,Oki B M,et al.Using collaborative filtering to weave an information tapestry[J].Communications of the Acm,1992,35(12):61-70.
[4]Resnick P,Iacovou N,Suchak M,et al.Group Lens:an open architecture for collaborative filtering of netnews[C].ACM Conference on Computer Supported Cooperative Work.ACM,1994:175-186.
[5]李灿,李书琴,蔡骋,等.基于IALM和填充可信度的协同过滤算法及其并行化研究[J].计算机应用研究,2016,33(10):2954-2958.
[6]Liu Z.Collaborative Filtering Recommendation Algorithm Based on User Interests[J].International Journal of u-and e-Service,Science and Technology,2015,8(4):311-320.
[7]孙光福,吴乐,刘淇,等.基于时序行为的协同过滤推荐算法[J].软件学报,2013,24(11):2721-2733.
[8]王骏,王士同,邓赵红.聚类分析研究中的若干问题[J].控制与决策,2012,27(3):321-328.
[9]Choi k,Suh Y.A New Similarity Function for Selecting Neighbors for Each Target Item in Collaborative Filtering[J].knowledge-Based Systems,2013(37):146-153.
[10]kaleli C.An Entropy-Based Neighbor Selection Approach for Collaborative Filtering[J].knowledge-Based Systems,2014(56):273-280.
[11]袁正午,陈然,李建华.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.
[12]谢毅刚,郭卫斌,李建华.基于双层相似度的协同过滤推荐算法[J].华东理工大学学报,2018,2(2):131-137.
[13]吴毅涛,张兴明,王兴茂,等.基于用户模糊相似度的协同过滤算法[J].通信学报,2016,37(1):198-206.
[2]荣辉桂,火生旭,胡春华,等.基于用户相似度的协同过滤推荐算法[J].通信学报,2014,35(2):16-24.
[3]Goldberg D,Oki B M,Oki B M,et al.Using collaborative filtering to weave an information tapestry[J].Communications of the Acm,1992,35(12):61-70.
[4]Resnick P,Iacovou N,Suchak M,et al.Group Lens:an open architecture for collaborative filtering of netnews[C].ACM Conference on Computer Supported Cooperative Work.ACM,1994:175-186.
[5]李灿,李书琴,蔡骋,等.基于IALM和填充可信度的协同过滤算法及其并行化研究[J].计算机应用研究,2016,33(10):2954-2958.
[6]Liu Z.Collaborative Filtering Recommendation Algorithm Based on User Interests[J].International Journal of u-and e-Service,Science and Technology,2015,8(4):311-320.
[7]孙光福,吴乐,刘淇,等.基于时序行为的协同过滤推荐算法[J].软件学报,2013,24(11):2721-2733.
[8]王骏,王士同,邓赵红.聚类分析研究中的若干问题[J].控制与决策,2012,27(3):321-328.
[9]Choi k,Suh Y.A New Similarity Function for Selecting Neighbors for Each Target Item in Collaborative Filtering[J].knowledge-Based Systems,2013(37):146-153.
[10]kaleli C.An Entropy-Based Neighbor Selection Approach for Collaborative Filtering[J].knowledge-Based Systems,2014(56):273-280.
[11]袁正午,陈然,李建华.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.
[12]谢毅刚,郭卫斌,李建华.基于双层相似度的协同过滤推荐算法[J].华东理工大学学报,2018,2(2):131-137.
[13]吴毅涛,张兴明,王兴茂,等.基于用户模糊相似度的协同过滤算法[J].通信学报,2016,37(1):198-206.