Trust Recommendation And Application In Community Education Curriculum

The rapid development of the Internet has brought massive data.In the face of this near disaster data,how to find the desired information has become a very important issue.In the traditional sense,the search engine does not consider the personalized needs of users,but the trust recommendation algorithm under collaborative filtering takes the personalized needs of users into account,especially the hybrid recommendation system formed by the superposition of various algorithms,which can learn from each other's strengths and make up for the weaknesses,reduce the shortcomings of a single algorithm,and play its advantages.This paper focuses on the relationship of trust and dynamic recommendation under the relationship of trust,and further applies the research to community education curriculum recommendation.The main research work and innovations are as follows:Based on the classic Plackett-Luce model of list level sorting learning,and integrating personalized network recommendation,L~2R~2SN algorithm is proposed.L2r2sn algorithm digs out the mutual influence factors and hidden features of the project from the friend's interaction network,integrates them into the Plackett-Luce model,uses the gradient descent method to iteratively optimize,trains the optimal solution of the model,selects the priority recommendation items according to the project sorting list,and recommends them to the users.Experimental results show that L~2R~2SN algorithm improves the accuracy of recommendation and reflects users'preferences more effectively.In order to solve the problem of cold start of dynamic recommendation system,a method of modeling user trust relationship and user rating data is proposed.The experimental results show that the method based on trust dynamic recommendation has better performance than the existing algorithm,can improve the accuracy and coverage,and better mine the long tail distribution of projects.It has been applied to the information dynamic recommendation of open education distance education base project construction,and has received good results.In view of the existing course recommendation algorithm is to solve the recommendation problem unilaterally from the course characteristics or user relationship,inspired by the principle of computer cache,this paper proposes a professional course pool cache algorithm which integrates trust relationship and random walk algorithm.According to the characteristics of the same professional courses,we use the similarity to classify the courses in the system and build the professional course pool cache.Then we consider the trust relationship and select the trust set.Finally,we use the average score of the same professional courses in the cache or the score of the most trusted users or the average score of the most trusted users in the same professional course pool cache as the prediction value.Experiments show that the accuracy of this recommendation is very high in a small community education system.