Research On E-Learning Recommendation Algorithm Based On Learning Style

At present,the personalized recommendation algorithm has been widely used in life,such as product recommendation and music recommendation of the e-commerce platform.The collaborative filtering recommendation algorithm is favored by researchers because of its simplicity and efficiency,but at the same time,there are also some deficiencies such as cold start and data sparsity.In order to optimize the recommendation results of collaborative filtering recommendation algorithm in E-Learning application domain,this paper improves the user-based collaborative filtering recommendation algorithm by doing much work on in-depth analysis of the unique rules and uniqueness of the domain to provide more accurate learning resources for learners.Firstly,a deep research is made on the personalized recommendation algorithm in E-Learning application domain,and it is found that the learner's learning style is one of the important reasons that have an impact on personalized recommendation.To solve the problems of high cost,low efficiency,and high subjectivity because of the traditional learning style prediction model's high dependence on the learning style scale,this paper proposes a learning style prediction method based on the implicit feedback information of the learner with the unsupervised learning algorithm(Learning Style Mining for short).And it uses improved K-means clustering and other related algorithms to predict learner's learning style.Secondly,in order to optimize the personalized recommendation results of the personalized recommendation system in the E-Learning application domain,this paper proposes a collaborative filtering recommendation algorithm adding the learning style data.The algorithm first uses implicit feedback information data to analyze learners' trajectories to establish an implicit scoring matrix model,and then uses the predicted learner's learning style as a reward factor for the similarity between learners to improve the traditional Pearson similarity algorithm which calculates the nearest neighbor set of the target learnerand gives a personalized recommendation for the learner with the learning resources to help the learner learn better.Finally,the improved algorithm proposed in this paper is verified on the real experimental data set,which implies that it is better than the original algorithm.