Research On Context-Aware Information Collaborative Filtering Recommendation Algorithm Based On Spark

With the development of information technology,the Internet and the explosive growth of data,users are increasingly engaged in online shopping,learning,social networking,and entertainment activities.The selection of information has been becoming a problem.The problem that needs to be solved immediately is how to help the user find the corresponding information.In many ways,the recommendation algorithm is one of the best representatives.On the one hand,traditional recommendation algorithms still have shortcomings,such as sparse data,cold start etc.,which leads to unsatisfactory results for users.On the other hand,algorithms cannot always recommend hot resources for users.Therefore,considering a variety of factors can help the user find the most appropriate and user-friendly information.The traditional recommendation algorithms only consider the similarity between users and projects,and do not consider the context-aware information of the data.Researches show that this information affects user preferences and choices.First,this paper analyzes the current problems of collaborative filtering algorithms,such as cold start,sparse data,and low accuracy.Aiming at the existing problems,a new collaborative context-aware collaborative filtering algorithm is proposed.The context-aware information of the combined data is used in the calculation process.The optimized algorithm has the following advantages:On the one hand,context-aware information of the data can be found,and on the other hand,using context-aware information for calculation can improve accuracy.Experimental verification of the optimized collaborative filtering algorithm.Analysis and comparison of experimental results show that the optimized algorithm can improve the accuracy of the recommendation algorithm.Secondly,we study the model-based matrix decomposition collaborative filtering algorithm in distributed environment.In the calculation process,data context-awareness information is combined,and the Spark platform has the advantages of rapid iterative calculation.Experimental verification of the optimized collaborative filtering algorithm.Analysis and comparison of experimental results prove the effectiveness of the optimized algorithm.In this paper,the proposed algorithm is used to perform multiple experiments on multiple data sets.The experimental results show that the optimized collaborative filtering algorithm achieves good results in terms of accuracy and RMS error evaluation.