A Design Of Correlation Algorithm Based On Data Mining

A wild progress of computer technology and network communication comes up due to thecurrent social and economical development. It’s faster, easier and cheaper for people to acquiredata with the help of growing information technology, which makes data and its informationincrease exponentially. People feel severe stress of "information explosion","chaoticinformation space","data glut" and "data tomb" confronted with such extremely expandingdata.Through study and borrowing of the advanced theoretical research achievement andexperience of domestic and overseas scholars, this paper considers dividing data into batchesaccording to the idea of Cloud Computing that processes input of data set in batches and step bystep. At first, estimate the result of first part data processing. At the same time, learn newknowledge and process data of the second part; then, take copular contiguous function astheoretical basis, design correlation measure contiguous function, and analyze the estimatedresult of first-part data obtained and the newly learnt knowledge of the second-part data jointly;through correcting the knowledge obtained, estimate a more correct correlation measure, thusgive out a step-by-step measure algorithm that effectively correlates mass data. Feasibility ofthis algorithm has been verified through simulation experiment. Result shows that thecorrelation algorithm designed in this paper can noticeably improve efficiency of correlationeffect measurement, and can effectively solve measurement problem of correlation effect ofsuper mass data or even unlimited data, which provides a thought reference for the developmentof correlation measuring tool during the Cloud Computing era.There is not a comprehensive presentation to data mining from the perspective of computeralgorithm, but just focuses on correlation analysis in the mass of data and network technologyenvironment from the statistical viewpoint in this article.