Research And Implementation Of Online Learning Recommendation Algorithm For Rating Prediction And Learning Style Filtering

Recommendation algorithms play an important role in solving the individual needs of users.Online learning recommendation algorithm not only needs to accurately recommend the user's preferred courses,but also needs the recommended courses to adapt to the user's learning style.In order to provide users with an efficient personalized course recommendation service,this paper conducts the following research work:(1)In order to accurately recommend the user's preferred courses,this paper proposes an SVD++?Time score prediction recommendation algorithm incorporating time effects.In order to select the best-performing scoring prediction recommendation algorithm under online learning data,four common recommendation algorithms are analyzed and studied experimentally.The experimental research shows that the SVD++recommendation algorithm is the best.However,due to the lack of a time factor in the SVD++ recommendation algorithm,it cannot solve the problem of changes in user course preferences caused by changes in time.Therefore,this paper proposes a SVD++?Time recommendation algorithm that integrates time effects to solve the problem of changes in user interest in courses under the influence of time.Experimental research shows that the algorithm effectively improves the accuracy of online learning score prediction.(2)In order to ensure that the recommended courses have better adaptability to the user's learning style,this paper proposes an online learning recommendation algorithm with learning style filtering.Because the recommendation algorithm based on score prediction alone cannot solve the possible inadaptability between the recommended course and the user's learning style,this paper introduces the Felder-Silverman learning style theory of pedagogy,and explicitly judges the learning style of different users based on the online ILS learning style scale..However,due to the insignificant learning style discrimination problem in the original learning style generation algorithm,the learning styles of different users cannot be distinguished efficiently.Therefore,this paper improves the learning style generation algorithm,and experimental research shows that it effectively improves the discrimination of learning styles.In order to facilitate the system to implement the learning style filtering algorithm,a label model is introduced,and on the basis of the course list recommended by the scoring prediction algorithm,an online learning recommendation algorithm for learning style filtering is constructed.(3)In order to accurately recommend users' preferred courses and ensure that the recommended courses are well adaptable to their learning styles,this paper designs and implements an online learning recommendation system based on an online learning recommendation algorithm based on score prediction and learning style filtering.In order to solve the problem of large-scale data processing and calculation in the online learning recommendation system,this paper uses Spark as the core computing engine,uses the popular distributed document database Mongo DB to save data,and uses Spring Boot technology as the main framework of the system.The online learning recommendation system designed and implemented in this paper provides users with efficient personalized course recommendation services.