The content personalization model based on collaborative filtering

V. Mitroshyn N. Shapovalova І. Dotsenko N. Saitgareev
Received 25.02.2021, Revised 04.04.2021, Accepted 27.04.2021
Abstract

Purpose. To develop and theoretically substantiate the efficiency of a personalized recommendation system of goods, services, or content that is based on machine learning technology, which combines such approaches to the recommendations personalization as collaborative filtering and content filtering. Develop a mechanism for determining the expediency of using certain metrics to detect user similarities and implement it in real-time on the development system. Create a secure service with a recommendation-personalized system for goods, services, or content, ensuring the protection of the personal users' data. Research methods. The following research methods are used: analysis of sources on the research topic, metric methods for determining the affiliation of objects to a certain group by their similarity, methods of artificial intelligence theory, modeling the learning process of classification algorithms, formalization of models, software design methods for software development, empirical methods of substantiation of optimal parameters of model learning, methods of object-oriented design and programming. Scientific novelty. The developed model of the recommendation system allows at the initial stage of the system's operation to assess the user's preferences, almost without knowing the information about his preferences. The paper proposes a mechanism for selecting a metric to determine the group of user affiliation, which should ensure the accuracy of the recommendation. Practical significance. A developed web service that provides personalized recommendations for goods, services, or content on the user's preferences, ensuring the protection of personal information and high accuracy of the recommendations made. The recommendation system module is designed in the form of an application software interface that allows its use on any web platform. Results. A personality-oriented approach to product, service, or content recommendations is proposed, a universal recommendation system has been developed, which combines both collaborative and content filtering, and web service has been developed, taking into account the security requirements of users' personal data

Keywords:
recommendation system; machine learning; metric method; metrics; collaborative filtering; content filtering

Retrieved from Vol. 19, No. 1, 2021

Pages 142–146

References Suggested citation

[1] The Netflix Prize Rules [Електронний ресурс] / – 2006. – Режим доступу до ресурсу: https://www.netflixprize.com/assets/rules.pdf.

[2] Stephanie Crawford. How Netflix Works [Електронний ресурс] / Stephanie Crawford, Tracy V. Wilson. – 2010. – Режим доступу до ресурсу: https://electronics.howstuffworks.com/netflix2.htm.

[3] Michael Jahrer. The BigChaos Solution to the Netflix Prize 2008 [Електронний ресурс] / Michael Jahrer, Andreas T¨oscher. – 2008. – Режим доступу до ресурсу: https://www.netflixprize.com/assets/ProgressPrize2008_BigChaos.pdf.

[4] Bharat Bhasker. Recommender Systems in e-Commerce / Bharat Bhasker, K Srikumar. // EC '99: Proceedings of the 1st ACM conference on Electronic commerce. – 2010. – С. 158–166.

[5] Fleder D., Hosanagar K. Blockbuster Culture's Next Rise or Fall: The Impact of Recommender Systems on Sales Diversity (журнал) // Management Science, Vol. 55, No. 5, May 2009, pp. 697-712. — 2009. — P. 1 - 49.

[6] Muffaddal Qutbuddin Comprehensive Guide on Item Based Collaborative Filtering [Електронний ресурс] / Muffaddal. – 2020. – Режим доступу до ресурсу: https://towardsdatascience.com/comprehensive-guide-on-item-basedrecommendation-systems-d67e40e2b75d.

[7] Гудфеллоу Я., Бенджио И., Курвилль А. Глубокое обучение / пер. с анг. А. А. Слинкина. – 2-е изд., испр. – М.: ДМК Пресс, 2018. – 652 с.: цв. ил.

[8] Сегаран Т. Программируем коллективный разум. – Пер. с англ. – СПб.: Символ-Плюс. 2008. – 368 с.

[9] Введение в кросс-валидацию k-fold [Електронний ресурс] – Режим доступу до ресурсу: https://codecamp.ru/blog/cross-validation-k-fold/.

[10] Introduction to scikit-learn [Електронний ресурс] - Режим доступу до ресурсу: https://neurohive.io/ru/osnovydata-science/vvedenie-v-scikit-learn/

[11] Сайт компанії IMDb – Режим доступу до ресурсу: https://www.imdb.com/

[12] Kaggle [Електронний рсурс] / Режим доступу до ресурсу: https://www.kaggle.com/ashirwadsangwan/imdbdataset

[13] Luc Bouganim, Yanli Guo. Database Encryption // Encyclopedia of Cryptography and Security / Ed. by Henk C. A. van Tilborg and Sushil Jajodia. — Springer, 2011. — P. 307—312.

[14] Rakesh Agrawal, Jerry Kiernan, Ramakrishnan Srikant, Yirong Xu. Order preserving encryption for numeric data (англ.) // Proceedings of the 2004 ACM SIGMOD international conference on Management of data - SIGMOD '04. — Paris, France: ACM Press, 2004. — P. 563

Mitroshyn, V., Shapovalova, N., Dotsenko, І., & Saitgareev, N. (2021). The content personalization model based on collaborative filtering. Journal of Kryvyi Rih National University, 19(1), 142-146. https://doi.org/10.31721/2306-5451-2021-1-52-142-146
en