Design And Implementation Of Production Process-oriented Isonicotinic Acid Yield Prediction System

During this period,following the dramatic breakthroughs in big data technology and intelligent optimization algorithms and advanced studies,the value of industrial production data has attracted a great deal of attention.As an important medicine mesocosm,it is used in the manufacture of anti-tuberculosis treatment,as well as an anti-corrosive agent and an additive for galvanizing,etc.It has a broad application prospect.In response to the problems of yield prediction relying on manual experience and low accuracy of prediction in isonicotinic acid production,this paper builds a visualization system using isonicotinic acid production data to predict yield,which can improve production efficiency and optimize the production process.The main work of this paper is as follow:(1)In this paper,based on the isonicotinic acid production dataset,the missing values and outliers of the dataset are processed,and the feature reduction is performed by combining the production process and correlation analysis,and the feature selection is performed by random forest feature importance.(2)An improved GWO-BP network model is proposed for the problem that the classical BP neural network,it is easy to slip into partial minima in the train-up session.which leads to the inability to reach the global optimal solution,thus affecting its accuracy and generalization ability.The essential principle is that the positional of the grey wolf algorithm works as the values of the weights and triggers of the BP neural network.As the locations of the gray wolf keep evolving,the weights and triggers of the BP neural network algorithm are frequently revised,and the preferable position of the gray wolf is the optimal answer to the BP neural network,and the GWO-BP neural network model is eventually established to forecast the yield by isonicotinic acid.(3)Through comparative experiments,the three isonicotinic acid prediction models,GWO-BP,GA-BP and BP,were compared and analyzed,and the GWO-BP model wasselected as the prediction model of this system.(4)Based on the requirement analysis,the structure and functional model of the system was designed.Spring Boot framework was used to develop the back-end business functions of the system,Vue was used as the front-end framework to develop the system,Element UI was developed componentively,Redis was used as the cache of the system,Mybatis-plus was used for data persistence,and the trained prediction model in Matlab was called in Java was solved in the yield prediction module of the system.Finally,the isonicotinic acid yield prediction system was developed and implemented.