Research And Implementation On Learning Resource Recommendation For Online Education

With the rapid development of Internet applications and the exponential growth of the volume of online information,online education resources are becoming more and more abundant,which,on the other hand,has brought difficulties for the individuation.Therefore,the precise positioning of resources for users and the individuation of online education services has become a common concern of many experts and scholars.This thesis does some research and improvement on the traditional recommender algorithm,and focuses on corresponding solutions to the problems such as cold-start problem and sparsity problem.After that,the improved recommender algorithm is applied to the acquisition of learning resources in online education.In the end,a personalized learning resource recommender system is designed and implemented to optimize the learner’s online learning experience.The research works completed are as follows:1.This thesis introduces the domestic and foreign research status of online education resources and recommender systems,and expounds the research background of this thesis,and discusses the practical use of the resources recommender for learners.2.The history of the development of the recommender system is summarized.The concept of the experimental method and the evaluation are introduced.This part introduced the recommendation algorithm in detail,and compared the differences between different methods.Afterwards,The concepts of learning resources such as organization form,neural network and so on are introduced.3.A collaborative filtering recommender algorithm based on Two-Attribute rating Matrix and neural network is proposed.This thesis uses users’ attributes and projects’ attributes to establish a Two-Attribute rating Matrix,and uses it to measure the similarity of users and establish the user attribute preference model to predict the rating of unkown items which is effectively mitigating the sparsity problem.When the new user or item enters the recommender system,their attributes are matched with the Two-Attribute rating Matrix,and the items or users with a good rating are selected to the users.This method which can solve the cold-start problem is effective.Experiments show that the proposed algorithm is effectively improving the quality of the recommender algorithm,especially for the new item and new user.4.This thesis applies the optimized algorithm to the acquisition of learning resources in online education,including the standardization of learning resources and learner information modeling,the usage of BP neural network algorithm to train the learner’s attribute preference model.At last,the results of the optimized recommender algorithm is compared and evaluated with the results of the traditional collaborative filtering recommender algorithm.5.A personalized learning resource recommender system based on online education is constructed and each module of the system is designed and implemented.The development platform and tools of the system are described and the actual development effects are tested.Through the integration of the recommender system,learners get a better personalized resource recommender service.The learning efficiency and learning experience are improved,laying the foundation for further research and application in the future.