Construction of personalized recommendation model for educational video game resources based on knowledge graph

Abstract

With the development of education and technology, teachers have gradually realized that games should not be just a way for students to entertain themselves. Applying games to teaching resources can achieve better teaching outcomes. However, related resources are constantly emerging on the internet. To achieve higher quality recommendations, a personalized recommendation model for educational video game resources based on knowledge graphs is proposed. Firstly, feature extraction is performed alternately on the user side and the item side. Then a hidden Markov model is introduced on the basis of the dual end neighbor algorithm. Considering the temporal nature of the user, the model is optimized. The optimized model takes into account the long-term and short-term preferences of users and mines their potential preferences. Through experimental analysis, the hit rate index value of the designed model reaches 0.7989. The normalized cumulative gain value of the broken line is 0.6045. More than 89 % of users are satisfied with the recommendation of this model. The running time is 0.2863 s. The constructed model can achieve efficient and high-quality recommendation of educational video game resources, providing users with a more convenient and efficient online experience.