抵抗素定量检测体系的回归曲线的选择及回归方程的建立

The choice of regression curve and establishing of regression equation in the quantitative analysis system of resistin

目的选择抵抗素的定量检测体系的最佳回归曲线和建立回归方程。方法1.运用ELISA原理及方法进行实验操作,检测加入不同浓度的抵抗素的定值所对应的吸光度,每个观测值进行三次重复平行测量,并记录实验结果,取其均值。2.采用直线回归方程、二次方程、三次方程、复合曲线、复合比级数曲线、指数方程、Logistic曲线、乘幂曲线、S形曲线方法分别对抵抗素的加入的定值与其对应的吸光度进行曲线拟合。3.据决定系数选择最佳回归方程。结果1.所采用的直线回归方程、二次方程、三次方程、复合曲线、复合比级数曲线、指数方程、Logistic曲线方法能进行曲线拟合;乘幂曲线、S形曲线方法不能进行曲线拟合。2."1"中各种拟合曲线的决定系数分别0.95908、0.98946、0.99992、0.78277、0.78277、0.78277、0.782779。结论1.根据上述抵抗素各回归方程的复相关系数、决定系数、校正复相关系数、标准误、回归方程及偏回归系数的假设检验,认为三次方程能较好地体现自变量与应变量的相关关系、曲线拟合也是最好的。2.抵抗素的最佳回归方程为:y=84.30+62.24x+8.69x2-0.31x3。

Objective To choose the best regression curve equation in the quantitative analysis system of resistin and to establish the regression curve equation.Methods 1.The corresponding absorbency of the resistin fixed value in different concentration was detected by ELISA theory and each observation value was measured for three repetitive parallel times.The experimental results were recorded and averaged.2.The corresponding absorbency was curve fitted to the resistin fixed value by Linear,Quadrati,Cubic,Compound,Growth,Exponent,Logistic,Power and S curve.3.The best regression curve equation was chosed according to determination coefficient.Results 1.It could be curve fitted by Linear,Quadrati,Cubic,Compound,Growth,Exponent and Logistic.It could not be curve fitted by Power and S.2.The determination coefficients of the above-mentioned fitting curve were 0.9591,0.9895,0.9999,0.7828,0.7828,0.7828 and 0.7828 respectively.Conclusion 1.It was thought that cubic equation was better to reflect the correlativity between independent variable and dependent variable and that it was best curve fitted by cubic equation according to the hypothesis tests of coefficient of multiple correlation,coefficient of determination,adjusted plurality correlation coefficient,standard error,regression equation and partical regression coefficient in the above-mentioned regression equations of resistin.2.The best regression curve equation of resistin was y=84.30+62.24x+8.69x2-0.31x3.