Research And Implementation Of Personalized Recommendation Course Algorithm For Online Education

Online education is a new model of education using advanced information technology to transfer the high quality teaching resources from the offline to online, with which learners can study through the modern network. With excellent Internet learning environment and high-speed network transmission, learners can learn anytime and anywhere. However, with the increase of network teaching resources, how to choose the suitable courses has become one of the problems facing by learners. Although the search engine can resolve the learner's requests in some cases, it is difficult for search engines to find the corresponding teaching resources when learners can hardly transform their learning requirements into search keywords. Personalized recommendation system can solve this problem well. The recommendation system can recommend the relevant teaching resources to learners by analyzing their history data and establishing their interest model.This paper proposes a personalized recommendation method for online education. The method improves the user-based collaborative filtering recommendation algorithm by analyzing and studying the features of online education. Aiming at the data sparseness and a small number of user ratings, this method optimizes the user similarity calculation module in the process of recommendation. The personalized recommendation algorithm optimizes the user similarity calculation module in recommendation process proposed, by introducing a new calculation factor and object properties. It can effectively reduce the similarity between "effective user" and "noise user", and reach the purpose of improving the accuracy of the recommendation. It can be applied to the weights of artificial and machine recommended courses, and the aim of improving the accuracy of recommendation can be achieved. The recommendation system based on collaborative filtering recommendation algorithm is implemented in the Java platform to deploy offline experiments. The results show that the personalized recommendation algorithm presented in this paper is better than the original algorithm in the online education data.