摘要
研究了一种基于颜色读数的物质浓度测量模型.首先分析几种物质在不同浓度下的颜色读数,构建多元线性模型,并通过逐步回归法解决了多重共线问题;降维后在Matlab中构建非线性模型,比较两种模型的优劣,结果表明非线性模型优于线性模型.
With the improvement of photographic technology and color resolution,it is helpful for establishing a quantitative relationship between color readings and substance concentration,that is,the concentration of the substance to be measured can be obtained by inputting the color readings in the photograph. Firstly,a multivariate linear model is constructed,and the correlation coefficients of the model are obtained by multivariate linear regression analysis. The multicollinearity problem is solved by stepwise regression method. After dimensionality reduction,a nonlinear model is constructed in MATLAB and the correlation coefficients are obtained. Finally,RMSE is used to analyze the model errors and compare the advantages and disadvantages of the two models.
引文
[1]王雪萍.颜色读数与物质浓度关系的数学建模[J].经济数学,2018,35(2):105-110.
[2]全国大学生数学建模竞赛.颜色与物质浓度辨识[EB/OL].[2017-9-14].http://www.mcm.edu.cn/html-cn/.
[3]乐薇,刘乐.数码成像比色法测定牛乳中蛋白质的含量[J].食品科学,2016,37(4):154-157.
[4]刘玲玲,薛澄岐.面向多维KE模型构建的产品特征解构方法[J].工程设计学报,2014,21(4):323-328.
[5]赵松山.对多重共线性的深入思考[J].当代财经,2003(6):125-128.