A METHOD OF AUTOMATED GENERATION OF RECOMMENDATIONS FOR EDUCATION APPLICANTS BASED ON A HYBRID APPROACH

DOI: 10.31673/2412-4338.2023.042127

Abstract

The most popular areas of use of recommender systems are platforms containing movies, music, books, and other objects that are fairly similar in their characteristics. Quite a lot of both general and specific methods of building recommendations have been developed for such systems. The article discusses the method of automated generation of recommendations for education applicants based on a hybrid approach, which involves taking into account the content of the objects of interest of the student. An analysis of the existing approaches to the construction of the recommendation was carried out and key problems were identified. It was determined that taking into account the content of objects of interest and users allows to increase the accuracy of the recommendation. The article provides typical types of recommendations that students need in the educational process. Sources for the formation of recommendations are described, in particular, explicit evaluations of objects of interest obtained on the basis of mandatory student surveys; implicit evaluations obtained on the basis of analysis of the behavior of users in the system and the content of objects of interest and users. The characteristics of the objects of interest were studied and it was determined that they necessarily have short or long text descriptions. The key stages of the method of automated formation of the recommendation are determined, taking into account the specifics of the domain of the subject area. In order to increase the accuracy of the recommendations, it is proposed to perform a preliminary selection of potentially interesting objects for the user based on Natural Language Processing methods before applying collaborative filtering. This will allow taking into account, among other things, the semantic proximity of the user's interests and objects in the system even with small amounts of textual information about objects of interest and users. In addition, the accuracy of the recommendation can be increased due to the use of explicit assessments of objects of interest, which are obtained at the level of the educational institution by regular measures to monitor the quality of education.

Keywords: recommendation system, collaborative filtering, hybrid approach, Natural Language Processing, information technology, database, automation, decision support.

References
1. Tkachenko, L. V., & Khmelnytska, O. S. (2021). Peculiarities of the implementation of distance learning in the educational process of a higher education institution. Pedagogy of creative personality formation in higher and secondary schools. No. 75, Vol. 3. P.91-96. Doi: https://doi.org/10.32840/1992-5786.2021.75-3.18
2. Nikiforova, T. D., Kapshuk, O. A., & Nechepurenko, D. (2022). Educational and organizational work of the dean's office under martial law. Materials of the II Forum of the academic community "Education in conditions of war: realities, challenges and ways of overcoming". P.41-43.
3. Patel, Dhruval & Patel, Foram & Chauhan, Uttam. (2023). Recommendation Systems: Types, Applications, and Challenges. 2210-142. 10.12785/ijcds/130168.
4. Meleshko E. V. Problems of modern recommendation systems and methods of their solution / E. V. Meleshko // Management, navigation and communication systems. - 2018. - Issue 4. - P. 120-124. - Access mode: http://nbuv.gov.ua/UJRN/suntz_2018_4_25
5. Beheshti, A., Yakhchi, S., Mousaeirad, S., Ghafari, S. M., Goluguri, S. R., & Edrisi, M. A. (2020). Towards cognitive recommender systems. Algorithms, 13(8), 176.
6. Raghuwanshi, Sandeep & Pateriya, R.. (2019). Collaborative Filtering Techniques in Recommendation Systems. 10.1007/978-981-13-6347-4_2.
7. Yuan, Zhenning & Lee, Jong & Zhang, Sai. (2021). Optimization of the Hybrid Movie Recommendation System Based on Weighted Classification and User Collaborative Filtering Algorithm. Complexity. 2021. 1-13. 10.1155/2021/4476560.
8. Zhang, Q., Lu, J., & Zhang, G. (2021). Recommender Systems in E-learning. Journal of Smart Environments and Green Computing, 1(2), 76-89.