The Research And Implementation Of Process Big Data Analysis Model Recommendation And Algorithm Parallel Technology

With the mature development of industrial big data acquisition technology,process big data analysis has been equipped with a complete data resource base.However,due to the complex type of process data,large volume,complex type of analysis model and low utilization rate,data analysis often leads to insufficient data value mining and low efficiency of analysis and calculation.Therefore,a complete set of platform supporting the process of big data analysis and high performance computing is urgently needed.Therefore,in view of the above problem,this thesis takes the process of big data analysis as the research object,designs and develops the process of big data analysis platform,and gets through the integration of data and analysis process.In order to solve the problem of low utilization rate of analysis model,the model recommendation method is proposed,and the parallel computing framework of big data is used to improve the operation efficiency of the algorithm.The main research contents of this thesis are as follows:(1)By summarizing and analyzing the characteristics of the process big data analysis scenario and process,the overall architecture design of the process big data analysis platform was completed,including the basic platform layer,the data layer,the modeling layer,the model layer and the application layer of the big data,and the functional modules of the process big data analysis platform were planned on this architecture.(2)To solve the problem of low utilization rate of analysis model,this thesis studied the method of recommendation of analysis model.Based on the classification management system of process big data analysis model,a multi-dimensional labeling system of process big data analysis model was constructed.And based on this labeling system,The recommendation method of big data analysis model based on content and collaborative filtering was proposed,and the effectiveness of the method was verified by an example.(3)To solve the problems of low computational efficiency of the algorithm,the mainstream parallel computing framework of big data was studied,and the parallel computing method of Spark+Flink for process big data analysis algorithm was proposed.And based on Flink framework,the parallel transformation of Apriori algorithm,which is commonly used for parameter recommendation analysis of big data in process,is completed.(4)On the basis of the overall framework and functional modules of the platform,the development of the process big data analysis platform is completed by using Java Web development technology.