Proceedings of the 2019 8th International Conference on Software and Computer Applications | 2019
Research on MOOC System Based on Bipartite Graph Context Collaborative Filtering Algorithm
Abstract
The MOOC platform is a good tool to help people learn, but with the increase of resources and numbers on the platform, choosing a learning resource that suits you has become a big problem. Personalized recommendations can help learners alleviate the problem, and recommending algorithms is the core of the recommendation process. Based on the analysis of the existing algorithms in the existing MOOC platform, in order to improve the accuracy and effectiveness of the recommendation and solve the cold start problem, this paper proposes a bipartite graph context collaborative filtering algorithm based on the characteristics of the MOOC platform: first, combined with the context information, the collaborative user filtering algorithm is used to process the initial user-resource score data and obtain the nearest neighbor. second, the nearest neighbor was used to get new user-resource score data. Last, in order to get a lists of recommendations, the bipartite graph method was used to process the new data. The algorithm improves the recommendation quality of the algorithm by preprocessing the data set and constructing different contact quantized values. Finally, the effectiveness of the algorithm is verified by experiments.