Research And Implementation Of Industrial Internet Big Analysis Platform

In recent years,with the deep integration of Internet technology and manufacturing industry,the realization of intelligent manufacturing has become an important step in the strategic development of all countries in the world.As a bridge connecting industry and the Internet,the industrial Internet platform provides a supporting carrier for the analysis of industrial big data.Although the research on the Internet platform in China continues to deepen,a unified standard for the framework of the industrial Internet platform has not been formed.Internet in our country,based on industry big data analysis platform for data collection and management technology has made rapid development,but the data search ability is weak,difficult to make full use of high value data,abandon the low value data complete industrial big data analysis,data analysis for industry has great influence on accuracy and efficiency.Therefore,this paper takes the industrial Internet big data analysis platform as the research object,studies the technology of data search,data processing,data parallelization analysis and so on,and designs and develops the industrial Internet big data analysis platform on this basis.The main research contents are as follows:(1)According to the design requirements of the big data analysis platform of the industrial Internet,the big data analysis framework of the industrial Internet is proposed.(2)To solve the problem of low data search efficiency of industrial Internet platform,a data search optimization method of "inverted index + industrial word segmentation" is proposed,which improves the data search efficiency of industrial Internet big data analysis platform.(3)In view of the problems of incomplete industrial high-quality data samples,many invalid data,and large data volume,a data processing and analysis scheme of "GAN-based data enlargement + LLE data dimension-reduction + parallel K-means" was proposed.Complete data simulation generated by GAN theory,solve the industrial problems such as large data samples is incomplete,unbalanced,LLE way to complete data dimension reduction,solve the industrial data invalid character too much,the problem such as low dimensional data distribution is not obvious,based on the Spark big data processing framework complete K-means clustering algorithm parallelization transformation,to solve the problem of low efficiency of industry analysis and calculation of large data,improve the efficiency of data processing and analysis.(4)Based on Hadoop,Spark and other open source big data frameworks,the development of the prototype platform for industrial Internet big data analysis was completed by using the mainstream Internet development technology of Java Web.