Hybrid Collaborative Filtering Recommendation of Educational Resources Based on Label Similarity and Learner Characteristics

At present, the recommendation systems of most online educational resources have the problem of poor recommendation quality and fail to meet the personalized needs of learners. Therefore, a hybrid collaborative filtering recommendation algorithm for educational resources based on label similarity and learner characteristics is proposed. Firstly, the recommendation technologies based on collaborative filtering and tag systems were analyzed respectively; Secondly, the set of students' preferred nearest neighbors is calculated by constructing the learner feature matrix, and then the similarity between different learners is calculated by using the label similarity algorithm. Finally, the score combination prediction is carried out by combining the above two recommendation factors to complete the final personalized recommendation. The simulation test results show that, compared with the existing two recommendation algorithms, the proposed hybrid collaborative filtering recommendation algorithm performs better in terms of accuracy and recall rate, and has a better recommendation function.