Analysis of problems of use of recommendation systems during application in Smart оbjects

DOI №______

Abstract

The article discusses the problems of using recommendation systems when used in SMART objects. Recommendation systems are programs that open up completely new possibilities and try to predict which SMART objects a user may like, having certain information about their profiles, including both in SMART RETAIL and in SMART HOUSE. It is shown that the use of recommendation systems has objective advantages, it allows you to increase the efficiency of staff in attracting and retaining customers, combine marketing with organizational and technical means, in turn, it allows to increase the productivity and profit of the enterprise. The consequences of introducing a system of recommendations as a modern intellectual technology are considered, including the emerging negative conditions for the effective implementation of this intellectual technology. Consider the features of the implementation of recommendation systems. Corresponding to the task, namely: to determine the most appropriate approach for application, which allows you to define objects without having any idea of what these objects are for use in various fields: in SMART RETAIL and in SMART HOUSE. The described approaches to developing recommendation systems, namely: content-based filtering and collaborative filtering. It is proved that it is most expedient to apply collaborative filtering methods. Consider three methods of collaborative filtering: the collective experience of the group; building a mathematical model; hybrid. The problems in the application of recommendation systems are analyzed, namely: sparseness of data, «cold start», scalability, synonyms, fraud, diversity. «White crow» application in social networks. It is determined that the collaborative filtering method recommends objects without any idea of what they are.

Keywords: recommendation systems; collaborative filtering.

References
1. Xiaoyuan Su, Taghi M. Khoshgoftaar. A Survey of Collaborative Filtering Techniques A Survey of Collaborative Filtering Techniques // Hindawi Publishing Corporation, Advances in Artificial Intelligence archive, USA. 2009.
2. An integrated approach to TV Recommendations by {{Не перекладено|TV Genius||en|TV Genius}}. Архів оригіналу за 6 червень 2012. Процитовано 15 квітня 2015 р.
3. Koren Y., Bell R., Volinsky C. Matrix Factorization Techniques for Recommender Systems // Computer // IEEE. Т. 42, № 8. С. 30–37.
4. Su X., Magyh T. Khoshgoftaar Survey of Collaborative Filtering Techniques // Advances in Artificial Intelligence. 2009.
5. A Survey of Collaborative Filtering Techniques, 2009, Р. 3.
6. Fleder D., Hosanagar K. Blockbuster Culture’s Next Rise or Fall: The Impact of Recommender Systems on Sales Diversity // Management Science, Vol. 55, No. 5, May 2009. Р. 697–712: журнал. 2009. P. 1–49.
7. Yehuda Koren. Factor in the Neighbors: Scalable and Accurate Collaborative Filtering // Yahoo! Research, Haifa. 2009. P. 1–11.
8. Sammut C., Webb J. (Eds.). Encyclopedia of Machine Learning. NY, USA // IBM T. J.Watson Research Center, 2010. Т. 1. Р. 829–838.
9. Linden G., Smith B., York J. Item-to-Item Collaborative Filtering // IEEE Internet Computing, Los Alamitos, CA USA. 2003. P. 76–80.