Educational and Competitive Algorithmic Platform Providing Personalized Recommendations
Mykola Hirnyi, Ye. V. Levus · 2024
Recommender systems have become integral to various domains, including e-commerce, social media, streaming services, and more recently, educational platforms. Given the complexity of the educational process, factors such as cold start problems, data sparsity, and limited computational resources present significant challenges. Previous research indicates that while the use of large language models addresses many of these problems, it has not been specifically focused on their educational value. ChatGPT capabilities are used to propose and develop a recommender system for use in our educational and competitive algorithmic platform. To collect the data, the platform was adapted to collect user feedback on given recommendations. The results demonstrate that the recommender system has a positive impact on user engagement leading to personalized learning. Specifically, there was an observed increase of 46% in the time spent reading pages and an overall extension of 5% in user session duration.