STUDY OF METHODS OF BUILDING RECOMMENDATION SYSTEM FOR SOLVING THE PROBLEM OF SELECTING THE MOST RELEVANT VIDEO WHEN CREATING VIRTUAL ART COMPOSITIONS
DOI:
https://doi.org/10.26906/SUNZ.2022.4.094Keywords:
collaborative filtering, knowledge-based method, deep neural network with immersion layers, extended reality, immersivenessAbstract
Topicality. Due to the growing digitization of art, the tasks of improving immersiveness during user interaction with extended reality art systems arise. Research methods. Collaborative Filtering by Matrix Factorization, a Knowledge-Based Method, Deep Neural Network with Immersion Layers. The purpose of the article: using models of recommendation systems built on different principles, conduct a number of computational experiments on model data and, comparing the results, check which of the existing approaches to building recommendation systems will show the best results when solving our problem -building a system for choosing a virtual art composition. The results obtained. The effectiveness of various methods of building recommendation systems for solving the problem of video selection in virtual art compositions is analyzed, taking into account explicit and implicit user feedback. It has been verified that the most effective approach using a hybrid model, which combines the method of collaborative filtering, a method based on knowledge and a deep neural network with immersion layers. It is proven that thanks to the mathematical apparatus of deep neural networks, it is possible to effectively solve the problem of video select ion in virtual art compositions, taking into account the user's preferences. Conclusion. The approach developed in the work can be used to improve immersiveness during user interaction with extended reality art systems.Downloads
References
F. Ye. Image Art Innovation based on Extended Reality Technology // 2022 7th IEEE International Conference on Data Science in Cyberspace (DSC), 2022, С. 584-587.
Michael D. Ekstrand. Collaborative Filtering Recommender Systems [Текст] / Michael D. Ekstrand, John T. Riedl, Joseph A. Konstan // University of Minnesota. – 2010. – С. 82–112.
Joonseok Lee. A Comparative Study of Collaborative Filtering Algorithms [Текст] / Joonseok Lee, Mingxuan Sun, Guy Lebanon // Cornell University. – 2012. – С. 2-3.
Badrul Sarwar. Item-Based Collaborative Filtering Recommendation Algorithms [Текст] / Badrul Sarwar, George Karypis, Joseph Konstan, John Riedl // University of Minnesota, Minneapolis. – 2001. – С. 2-3.
Francesco Ricci. Recommender Systems Handbook [Текст] / Francesco Ricci, Lior Rokach, Bracha Shapira, Paul B. Kantor // New York Dordrecht Heidelberg London, 2011. – С. 76-85.
Dietmar Jannac. Recommender Systems: An Introduction [Текст] / Dietmar Jannac, Markus Zanker, Alexander Felfernig, Gerhard Friedrich // Technische Universität Dortmund, Germany, 2010. – С. 49-59.
Peter Bostron. Comparison of User Based and Item Based Collaborative Filtering Recommendation Services [Текст] / Peter Bostron, Melker Fhilipson // SKOLAN FÖR DATAVETENSKAP OCH KOMMUNIKATION. – 2017. – С. 6-7.
