MODIFIED ALGORITHM OF COLLABORATIVE FILTERING FOR FORMING USER RECOMMENDATIONS
Background. In our online life, we get a lot of information, and more and more people don`t want to rummage impassable jungle of information. Each of us wants to quickly find what is looking for. Many sites such as YouTube, Facebook and Twitter have already had recommender system and many people have used it. Recommender systems are becoming more and more popular.
Objective. The base algorithm of collaborative filtering which is used in recommender system is considered. We are trying to find bottleneck problems of base algorithm of collaborative filtering to improve it and take a gain in time.
Methods. We have analyzed the base algorithm of collaborative filtering and have found bottleneck problem. The main runtime of the algorithm is concentrated to calculate user similarity. We calculated the average rating for object in cluster with weighting factor. We use two criterions to compare the base algorithm with the modified algorithm. First criterion is the algorithm runtime. Second criterion is amount of elementary permutations we have to do to get recommendations which are provided by the base algorithm of collaborative filtering. The main factors which influence the algorithm runtime of collaborative filtering are: number of users, amount of objects and percentage of filling.
Results. The modified algorithm of collaborative filtering was compared with the base algorithm of collaborative filtering by two criteria. The difference between the results of both algorithms does not exceed 5%. The modified algorithm works faster than the base algorithm. Furthermore, with increasing the number of users or amount of objects the runtime difference will increase. The results of research are presented in graphs.
Conclusions. We have analyzed the base algorithm of collaborative filtering and methods to improve it. We can conclude on the feasibility of the modified algorithm of collaborative filtering from the research. The modified method gives a great gain in time. If systems start to use this modified algorithm, this can solve the problem with the runtime of the algorithm of collaborative filtering and allows giving recommendations faster than the system which uses the base algorithm.
P. Melville et al., “Content-boosted collaborative filtering for improved recommendations”, in National Conference on Artificial Intelligence, Edmonton, Canada, 2016, pp. 187–192.
V. Srikar and R. Sudha, “Examining lists on twitter to uncover relationships between following, membership and subscription”, in Proc. 22nd Int. Conf. World Wide Web, Rio de Janeiro, Brazil, 2013 pp. 673–676. doi: 10.1145/2487788.2488019
P. Chatterjee et al., Advanced Multi-Criteria Decision Making for Addressing Complex Sustainability Issues. Hershey, PA: IGI Global, 2019.
H. Jafarkarimi et al., “A naive recommendation model for large databases”, Int. J. Inform. Educ. Technol., vol. 2, no. 3, pp. 216–219, 2012.
J.S. Breese et al., “Empirical analysis of predictive algorithms for collaborative filtering”, in Proc. 14th Conf. Uncertainty in Artificial Intelligence, San Francisco, CA, 1998, pp. 43–52.
F. Ricci et al., “Introduction to recommender systems handbook”, in Recommender Systems Handbook. Boston: Springer, 2011, pp. 1–35. doi: 10.1007/978-0-387-85820-3_1
B. Sarwar et al., “Item-based collaborative filtering recommendation algorithms,” in Proc. 10th Int. Conf. World Wide Web, Hong Kong, China, 2001, pp. 285–295.
X. He et al., “Neural collaborative filtering”, in Proc. 26th Int. Conf. World Wide Web, Perth, Australia, pp. 173–182.
Y. Koren and R. Bell, “Advances in collaborative filtering”, in Recommender Systems Handbook. Boston: Springer, 2015, pp. 77–117.
C. Yin et al., “Improved collaborative filtering recommendation algorithm based on differential privacy protection”, J. Supercomput., vol. 76, pp. 5161–5174, 2020. doi: 10.1007/s11227-019-02751-7
Р. Lytvak and I. Kochubey, “Application of MAI to solve the problem of choosing an algorithm of the system of forming proposals to bookstore users”, in Aktuelle Themen im Kontext der Entwicklung der Modernen Wissenschaften, Dresden, Germany, 2019, pp. 62–68.
G.H. Tzeng and J.J. Huang, Multiple Attribute Decision Making. Methods and applications. Boka Raton: CRC Press, 2011.
GOST Style Citations
- There are currently no refbacks.
Copyright (c) 2020 The Author(s)
This work is licensed under a Creative Commons Attribution 4.0 International License.