PROBLEMS OF MODERN RECOMMENDATION SYSTEMS AND METHODS OF THEIR SOLUTION

Authors

  • Yu. Meleshko

DOI:

https://doi.org/10.26906/SUNZ.2018.4.120

Keywords:

recommendation systems, collaborative filtering, content filtering, cold-start, continuous cold-start, filter bubble

Abstract

The subject matter of the article is the processes of building recommendation systems. The goal is to investigate the problems of modern recommendation systems and to find methods for their solution. The tasks to be solved are: to investigate the problems of modern recommendation systems, to carry out the comparative analysis of known methods of constructing recommendation systems in terms of the availability/absence of these problems, to investigate the existing methods for solving these problems. The following results were obtained: the main problems of modern recommendation systems are considered: user cold-start, item cold-start, user continuous cold-start, item continuous cold-start, filter bubble. The comparative analysis of known methods of constructing recommendation systems in terms of the availability/absence of these problems is carried out. The directions of further research for the development of methods for solving existing problems of a recommendation systems have been determined. Conclusions. The main problems of a recommendation systems are: user cold-start, item cold-start, user continuous cold-start, item continuous cold-start, filter bubble. To date, the cold-start problem has been practically solved by using contextual information and building hybrid recommendation systems. At the same time, cold-start problem became actual and for today is not completely solved. A promising direction for solving the problem of continuous cold-start is the use of artificial intelligence algorithms to adapt for possible changes in the characteristics of objects and user preferences. To solve the filter bubble problem, additional requirements should be applied to the formation of a list of recommendations. A promising direction for solving the filter bubble problem, as the study has shown, is to provide a list of recommendations with such properties of its elements as diversity, serendipity, novelty. At the same time, these are very subjective indicators for which there are no generally accepted metrics for measuring them and reliable methods for ensuring their implementation.

Downloads

References

Jones M. Recommender systems, Part 1. Introduction to approaches and algorithms. Learn about the concepts that underlie web recommendation engines / M. Jones – 2013. [Електронний ресурс] – Режим доступу: https://www.ibm.com/developerworks/opensource/library/os-recommender1/index.html?s_tact=105agx99&s_cmp=cp

Мелешко Є.В. Дослідження методів побудови рекомендаційних систем в мережі Інтернет / Є.В. Мелешко, Г.С. Семенов, В.Д. Хох. // Збірник наукових праць "Системи управління, навігації та зв’язку". Випуск 1(47). – Полтава: ПНТУ ім. Ю. Кондратюка. – 2018. – С. 131–136.

Сегаран Т. Программируем коллективный разум. – Пер. с англ. – СПб: Символ-Плюс, 2013. – 368 с.

Xiaoyuan Su and Taghi M. Khoshgoftaar A Survey of Collaborative Filtering Techniques A Survey of Collaborative Filtering Techniques // Hindawi Publishing Corporation, Advances in Artificial Intelligence archive, USA : журнал. — 2009. – P. 1 - 19.

Burke R. Hybrid Web Recommender Systems / Burke R.. // The Adaptive Web. Lecture Notes in Computer Science, vol 4321. Springer, Berlin, Heidelberg. – 2007. – С. 377–408.

Bernardi L., Kamps J., Kiseleva J., Mueller M.J.I. The Continuous Cold Start Problem in e-Commerce Recommender Systems. – 2015. [Електронний ресурс] – Режим доступу: https://www.researchgate.net/publication/280773072_The_Continuous_Cold_Start_Problem_in_e-Commerce_Recommender_Systems

Меньшикова Н.В. Обзор рекомендательных систем и возможностей учета контекста при формировании индивидуальных рекомендаций / Н. В. Меньшикова, И.В. Портнов, И.Е. Николаев. // ACADEMY. – 2016. – №6. – с. 20–22.

Амелькин С.А. Оценка эффективности рекомендательных систем // Труды 14-й Всероссийской научной конференции «Электронные библиотеки: перспективные методы и технологии, электронные коллекции» – RCDL-2012, Переславль-Залесский, 15-18 октября 2012 г. – С. 288-291.

Recommender Systems Handbook / Editors Francesco Ricci, Lior Rokach, Bracha Shapira, Paul B. Kantor. – 1st edition. – New York, NY, USA: Springer-Verlag New York, Inc., 2010. – 842 с.

Castells P., Vargas S., Wang J. Novelty and Diversity Metrics for Recommender Systems: Choice, Discovery and Relevance. [Електронний ресурс] – Режим доступу: http://ir.ii.uam.es/rim3/publications/ddr11.pdf.

Kotkov D., Konstan J.A., Zhao Q., Veijalainen J. Investigating Serendipity in Recommender Systems Based on Real User Feedback. [Електронний ресурс] – Режим доступу: https://www-users.cs.umn.edu/~zhaox331/papers/denis2018sac.pdf.

Kaminskas M., Bridge D. Measuring Surprise in Recommender Systems. [Електронний ресурс] – Режим доступу: https://www.insight-centre.org/sites/default/files/publications/recsys2014.pdf.

Castells P., Vargas S., Wang J. Novelty and Diversity Metrics for Recommender Systems: Choice, Discovery and Relevance / P. Castells, S. Vargas, J. Wang. – 2011. [Електронний ресурс] – Режим доступу: https://www.semanticscholar.org/paper/Novelty-and-Diversity-Metrics-for-Recommender-and-Castells-Vargas/4ec6bd672aaaa075b42a751099eb9317857e6e0c

Published

2018-09-12