Font Size: a A A

E-learning Resource Personalized Recommendation System Based On Collaborative Filtering Technology Research

Posted on:2016-12-21Degree:MasterType:Thesis
Country:ChinaCandidate:N ZhaoFull Text:PDF
GTID:2297330464962352Subject:Education Technology
Abstract/Summary:PDF Full Text Request
With the rapid increasement of information resources on the network, people are enjoying the convenience of information but also sufferring from the problem of information overload. It makes difficult for users to find the information resources to meet their needs in limite time. At present, domestic and foreign experts and scholars also presents a variety of personalized information solutions to solve the information overload problem.hoping to make personalized resources for the Internet users.This paper firstly introduces the basic concepts and features of E-learning,emphasizes the relationship of E-learning and traditional forms of learning. Analysed the Application condition of collaborative filtering recommendation technology and related technology theory. Through the study of the personalized recommendation model, intuitively shows the system in each part of the content, and specific descripted the steps of recommended learning resources and personalized recommendation system. introduces in detail of the principle of collaborative filtering recommendation, focusing on the analysis of the collaborative filtering recommendation algorithm, finding the nearest neighbor, similarity calculating method and the recommendation process, also describes the problems of collaborative filtering recommendation technology.Solve the problem of sparse data resulting in recommendation accuracy, this paper proposes that to make full use of the background information of the user and the user attributes, compute the similarity between users. The analysis of the traditional collaborative algorithm to calculate the similarity of the project mainly depends on the user rating data.Paper used genetic k-means algorithm to calculate the similarity of user attributes. firstly represents the user attributes with binary form which can be recognized by genetic algorithm. Then crossover and mutation user’s attributes, and utilizes the fitness function to calculate the best seed classification of user attribute, the best seeds became the best initial clustering center of K-means recommended steps. After Introducing several nearest neighbors to the user,use thesimilarity calculating method and the recommendation process of the filtering recom mendation algorithm, At the same time, described the deficiencies existing in collabo rative filtering recommendation technology.Finally, design the implementation of the ELPRS(the E-learning personalized recommendation system). Firstly, design the system function module, elaborate the system function demand in details, display the working principle of recommender systerm according to the work flow chart of the system.Focusing on the introduction of the important role in this optimization algorithm in recommendation work,describing the establishment and development of the system database environment configuration, shows the realization of part of the source code with the front desk system user interface.
Keywords/Search Tags:E-learning, Personalized recommendation, Collaborative filtering, Clustering algorithm, Genetic algorithm k-means
PDF Full Text Request
Related items