Research And Implementation Of Hybrid Recommendation Model In Learning Resources

The popularity of Internet and Big data makes online learning applications have been rapid development.Online learning platform is convenient for people to obtain knowledge resources,promote interest in learning and personalized,helps to improve the learning efficiency,but also the existence of resource overload problem,it has a very important significance of retrieval efficiency to recommend the resources which user interested quickly and accurately and to improve the user experience.Based on the comprehensive analysis of the online learning platform system framework and the characteristics of learning resources,this paper uses the hybrid model to carry on the similar calculation of resources,and then realizes the learning resource recommendation.Hybrid recommendation model is an effective method for multi dimension feature recommendation.Hybrid recommendation model relies on the support of big data,aiming at learning resources,this paper through the analysis of online learning platform,using the text data of resources and user behavior history data,using the implicit rating model and LDA topic model weighted model for user preferences,to solve the traditional collaborative filtering algorithm without scoring problems.The learning resource recommendation algorithm is realized by using the project based collaborative filtering algorithm and the parallel ALS-WR algorithm based on matrix decomposition.Finally,in order to verify the performance of this model,and provide examples of the recommendation algorithm in the big data environment,use Github log data source,using the parallel algorithm of Resource Recommendation Mahout experiments and analysis,and verified the validity of the model.