Research On Course Recommendation Algorithm Based On Features Information And User Behavior

Nowadays,online education is constantly innovating and developing vigorously under the opportunity of the rapid development of the Internet industry.At the same time,online education platforms have introduced new ideas,and users’ learning methods have changed,and users can directly learn courses online.However,when users obtain information from many courses on the platform,it is often difficult to find courses that meet the needs of users.The birth of the course recommendation system has solved the problem of the big explosion of course information.With the continuous innovation of technology,deep learning models have gradually been applied to the field of course recommendation.Course recommendation algorithm can provide satisfactory recommendation results to target users.With the continuous increase in the number of online users and courses,the course recommendation system exposes problems such as diversified user characteristics,complex user interests and preferences,and variability in user behavior,which reduces the recommendation accuracy of the recommendation algorithm.In order to solve the above problems,this paper analyzes and improves the collaborative filtering recommendation algorithm,and proposes three course recommendation algorithms.There are three parts of the research work:1.Aiming at the problem that traditional collaborative filtering algorithms ignore user characteristics and user interest preferences,it is difficult to obtain a course recommendation algorithm that combines user characteristics and interest clustering.The algorithm obtains course prediction scores according to the importance of user characteristics and different user preferences for course keywords.The TF-IDF algorithm is used to construct a user interest preference matrix,Canopy and K-means clustering algorithms are used to cluster users to obtain course prediction scores,and finally weighting two kinds of predictive scores to realize course recommendation.In the process of improvement,the problem of data sparsity is alleviated and the performance of the recommendation algorithm is improved.2.Aiming at the problem that the interest preferences of online learning users are in the process of dynamic changes,a course recommendation algorithm based on the dynamic changes of users’ interests is proposed.The algorithm further controls the decay speed of the time function by introducing the user interest change curve and the time attenuation factor.At the same time,the average scoring method is used to capture the user’s recent interest,so as to effectively grasp the change of the user’s interest preference.According to comparative experiments,the improved algorithm improves the accuracy of course recommendation on the basis of improving data sparsity and timeliness,and the recommended content satisfies users.3.Aiming at the problem that the personalized course recommendation algorithm does not fully consider the multi-feature attributes of users and courses,and the user’s historical course selection behavior,a course recommendation algorithm based on embedded features and user behavior is proposed.The algorithm introduces multiple characteristics of users and courses,uses LSTM model to process course name features,and establishes a recommendation model.At the same time,the course information collaboration model is constructed by using the user’s course selection behavior,and the two models are combined to realize Top-N recommendation.The simulation experiment results show that the improved algorithm recommended by this course performs well in algorithm performance and recommendation accuracy,and also proves the effectiveness and feasibility of the algorithm.The recommendation algorithm proposed in this paper can be effectively used for online users’ course recommendation.The algorithm recommendation accuracy and user satisfaction are high,and the time complexity is low.It can be applied to practice to further improve the effect of the course recommendation system.