Experimental Study On The Correlation Between Paper Printability And Printing Quality

For the main characteristics of printed paper,based on the regression of lasso and Gaussian process regression,a method for establishing the printability and quality prediction model of paper is proposed.The non-objective evaluation problems based on subjective evaluation and evaluation of individual performance indicators when evaluating printability and quality of printed products is solved.The paper has quantified the behavior and function rule of paper in printing,which has positive theoretical significance for the research of printability and the event of Printing Science and technology.The performance indexes of the paper and the color phase error,the gray level,the color purity,the color efficiency and the color intensity of the yellow,product and green ink in the color reproduction are measured by the collection of paper on the IGT printing press.The correlation analysis of the paper performance index is.carried out,and the linear correlation of the two variables is calculated by the correlation analysis,and the correlation of the two variables is calculated.Coefficient;using R language to analyze the performance index and color effect index by simple linear regression analysis,get correlation coefficient and analyze its influence degree;based on Lasso regression and Gauss process regression,the prediction model between paper performance index and color rendering effect index is established,and the advantages and disadvantages of two models are compared.The Lasso regression prediction model and the model are compared.The Gauss process regression forecasting model is optimized by the method based on genetic coding to automatically generate the eigenvalues.The prediction model of the Gauss process regression is applied to verify the accuracy of the actual printing data application model,and the regression prediction model is used to classify the paper,and the paper is classified and predicted.The results show that the residual square sum of the Lasso regression model is 5.106751,the residual square sum of the Gauss process regression prediction model is 4.535301,and the accuracy of the Gauss process regression prediction model is better.After optimization,the residual square sum of the Lasso regression prediction model is 4.731065,the accuracy is 7.4%,and the Gauss process regression prediction model is improved.The sum of squares of the model is 3.987901,the precision is 12.1%,the precision of the model is improved,and the accuracy of the Gauss process regression model is higher.The prediction of actual printing data and the paper classification prediction are all in the range of error allowed by the prediction of the actual printing data and the paper classification prediction.