SLOPE ONE COLLABORATIVE FILTERING BASED ON USER SIMILARITY
Abstract
To solve the problem of data sparsity in traditional collab-orative filtering algorithm and improve the accuracy of rec-ommendation algorithm, a Slope based on user similarity is proposed.
One Filling Score Matrix Collaborative Filtering Algo-rithms. The algorithm first uses the basic cosine similarity to calculate the similarity between users and generates the user similarity matrix, then fills the original user similarity matrix with the score predicted value of Slope One algorithm as the backfill value of the first n neighbors, and finally recommends the filled similarity matrix according to the traditional item-based collaborative filtering algorithm. Slope One collabora-tive filtering algorithm based on user similarity takes full ac-count of the similarity between users when filling the matrix, which makes the score prediction more accurate. The new al-gorithm alleviates the problem of data sparsity to a certain extent and improves the accuracy of the algorithm. The im-proved algorithm, classical collaborative filtering algorithm and even-weighted Slope One algorithm are tested on Movie Lens dataset. The results show that the Slope One collabora-tive filtering algorithm based on user similarity effectively al-leviates the problem of data sparsity and has better recom-mendation effect.
Downloads
References
[2]. Herlocker L, Konstan A, Borcher S A, et al. An algorith-mic framework for performing collaborative filtering [C]//Proc of, International ACM SIG IR Conference on Research and Development in Information Re-trieval.1999:230-237.
[3]. You Haipeng, Li Hui, Wang Yunmin, et al. An improved collaborative filtering recommendation algorithm comb-ing item clustering and slope one scheme [C]. Lecture Notes in Engineering & Computer Science, Vol 2215.2015:313-316.
[4]. You Haipeng, Li Hui, Wang Yunmin, et al. An improved collaborative filtering recommendation algorithm
combining item clustering and slope one scheme [C] Lec-ture Notes in Engineering & Computer Science, vol2215 2015:313-316.
[5]. Xiangdong, Qiu Zixian. Research on collaborative filter-ing algorithm based on slope one algorithm to improve score matrix filling [J/OL]. Computer application research, 2019 (05): 1-5 [2018-09-21].
[6]. LI Liang. Dong Yuxin, Zhao Chunhui, et al. Collaborative filtering recommendation algorithm combined with user trust [J] Journal of Chinese Computer Systems, 2017, 38 (5) 951-955
[7]. Gao Fenglong, Xing Chunxiao, Du Xiaoyong, etc. Collab-orative filtering algorithm based on matrix clustering [J]. Journal of Huazhong University of Science and Technol-ogy. Natural Science Edition, 2005, 33 (S1): 257-260 (Gao Fengrong, Xing Chunxiao, Du Xiaoyong, Wang Shan A collaborative filtering algorithm based on matrix clustering [J] Journal of Huazhong University of Science and Technology (Natural Science Edition), 2005, 33 (S1): 257-260)
[8]. Goldberg D, Nichols D, Oki B M, et al. Use collaborative filtering to weavr an information tapestry [J]. Communi-cations of the ACM. December, 1992.35 (12): 61-70
[9]. Dehghani Z, Reza S, Salwah S, et, al. A systematic review of scholar context aware recommender systems [J] Expert Syst Appl 2015 (42): 1743
[10]. Adomavicius G, Tuzhilin A Toward the next generation of recommender system: a survey of the state-of-art and pos-sible extensions [J].IEEE Trans on Knowledge & Data Engineering, 2005, 17 (6): 734-749
[11]. MinhA, Salakhutdinow R. Probability matrix factorization [C]//Advances in Neural Information Processing Systems 2007:1257
[12]. Rosin, Ouyang Yuanxin, Xiong Zhang, etc. optimize col-laborative filtering algorithm based on K-nearest neighbor through similarity support [J]. Journal of Computer Sci-ence, 2010, 33 (8): 1437-1445.
[13]. MovieLens_100k [DB/OL]. https://grouplens.org /da-tasets /movielens/.
