Big Data Processing Method Based On Grey Theory

With the rapid development of internet,especially in the 21 st century,mankind has entered the era of information explosion.It is difficult to find exactly what you want in the massive network information,and excessive information has become a burden.In this case,recommended system cames into being.Although a lot of scholars have done research and improvement on collaborative filtering recommendation algorithm,data sparsity problem can't be solved perfectly.In this paper,based on a large number of literature research,and previous studies,gray relational analysis method is introduced to alleviate the inaccuracy of big data recommendation algorithm.The gray relational clustering method is used to cluster the project to reduce the project space of the target user and alleviate the data sparsity problem.At the same time,it can reduce the searching time of the user's nearest neighbors within the same project cluster.Finally,the simulation results are given to verify the effectiveness of the proposed algorithm.The main work of the paper is as follows:Firstly,the paper summarizes the present situation,advantages and disadvantages of large data processing scheme,and proposes a large data processing scheme based on Grey theory.Secondly,the gray theory is applied to the similarity calculation of item clustering and neighbor selection.In bug data processing,the application of gray correlation analysis improves the accuracy of recommendation.Thirdly,in the process of project clustering,the gray relational theory is used to realize the clustering and grouping work in the project.After clustering,the characteristics of the same kind of projects are highlighted,and the differences among different categories are highlighted.Fourthly,the gray similarity degree is used to calculate the similarity between users and find the nearest neighbor of the target user.Based on gray theory,compared with the other algorithms,the recommended experimental simulation analysis verifies the effectiveness of the algorithm.