The Research And Implementation Of Data Mining Visualization Platform Based On Web

In the context of the era of big data led by science and technology,massive data has brought data analysis needs to various fields,and the technical threshold,workload,and tediousness of data mining are also increasing.Based on the above requirements,many enterprises combine visualization technology and data mining technology,and display the data mining process and results to users in an intuitive form to improve the efficiency,accuracy and effectiveness of data mining.However,the existing data mining visualization platform still has the following shortcomings:1)The execution process of the data mining model does not combine the workflow and task scheduling logic of the underlying computing framework of the system,and the calculation performance utilization rate needs to be improved;2)The lack of complete data mining work life cycle considerations make the optimization of data mining modeling difficult,cumbersome,and repetitive;3)Lack of work reports on the complete process of data mining,fail to summarize the data mining work,and provide an effective way to optimize and improve.In view of the above shortcomings,the research content of this article is as follows:1)Design and implement the execution flow of the data mining pipeline model in conjunction with the distributed data mining framework Spark workflow.This process provides the system with basic data mining algorithms,and the task scheduling logic based on the distributed framework provides model translation technology and model execution process for the pipeline model.2)Design and implement a data mining visualization system in conjunction with the complete life cycle of data mining.The system supports users to drag-and-drop build data mining pipeline models,provides various data mining algorithms WebAPI,configuration modules,and visual execution results and log modules to provide a visual working environment for the complete process of data mining.3)Design and implement a data mining visual report subsystem.This system is based on a componentized design and provides users with a variety of operator report templates according to the data mining operator data types.It supports users to combine system report templates to select models for data mining projects.Data and result data are independently edited to generate data mining reports.Based on the above research content,this paper designs and implements a Web-based data mining visualization platform.The platform is based on the Spark distributed framework to provide users with efficient data mining computing capabilities,provide drag-and-drop pipeline modeling,highly connect users to the data mining process,and provide independent editing and system template data mining report generation function to present the complete life cycle of data mining to users in a highly visual manner.This visualization platform is of great significance for improving the performance of data mining,reducing the difficulty,complexity and repetition of data mining,and reducing the difficulty of data mining learning.