甘蔗蔗汁品质性状的回归分析及模型  

Regression Analysis and Models of Quality Characteristics in Sugarcane Juice

作　　者：高三基[1] 陈如凯[1] 张华[1] 傅华英[1] 邓祖湖[1] 罗俊[1] 徐良年[1] 

机构地区：[1]福建农林大学/农业部甘蔗生理生态与遗传改良重点开放实验室,福州350002

出　　处：《植物遗传资源学报》2005年第1期84-87,90,共5页Journal of Plant Genetic Resources

基　　金：国家863计划项目(2002AA241031)国家科技攻关先导项目(2002RA544C)资助

摘　　要：本文对47份不同甘蔗基因型的蔗汁品质性状进行简单回归和逐步回归分析。结果表明,蔗汁蔗糖分(Suc)与锤度(BX)、温度校正后锤度(BX’)、蔗汁旋光读数(0Z)以及转光度(pol)之间均呈线性回归,简单回归模型均达极显著水平(P<0.01),决定系数(R2)分别为0.9393、0.9199、0.9861及0.9839。通过逐步回归分析和残差分析建立多重线性回归方程为Suc=0.05706+0.21488BX+0.181030Z,方差分析表明多重线性回归模型达极显著水平(P<0.01),决定系数为0.9931。t测验表明,蔗汁蔗糖分的模型预测值与实测值间的差异不显著,相对误差平均为0.49%。Quality characteristics of forty-seven different sugarcane genotypes were performed by simply and stepwise regression analysis. The results of simply regression displayed that relationships of sucrose content (Suc) with Brix (BX), Brix after temperature correction to 20℃ (BX'), 0Z value of international sugar scale (0Z), and polarization (pol) in juice were linear regression. The regression models were extremely significant (p <0.01) and the determination coefficients (R2) were 0.9393, 0.9199, 0.9861 and 0.9839, respectively. Mu-ti-indepents regression equation was demonstrated by stepwise regression and residual analysis, that is Suc = 0.05706+ 0.21488BX+0.181030Z. The model was extremely significant (p<0.01) by variance analysis and determination coefficient was 0.9931. Difference of prediction values of the model with true values was not significant by t-test and average relative error was 0.49 % .

关 键 词：品质性状 蔗汁 甘蔗 多重线性回归模型 逐步回归分析 线性回归方程 显著水平 方差分析表 温度校正 残差分析 相对误差 蔗糖分 基因型 转光度 t测验 实测值 预测值 锤度 平均 

分 类 号：S566.1[农业科学—作物学]