基于实验数据的精确数学回归模型设计

在生产和科学实验中,因素与指标的关系,即数学模型是很难精确确定的.对因素进行水平划分后,通过实验来寻求其数学模型,选取的数学模型主要是线性回归方程形式,其实验次数确定数学模型中待定系数的个数.因素越多、实验次数越多时,这种方法会造成分析模型与实际模型间的误差较大,甚至造成实际生产应用中不可估量的损失.本文通过对回归分析模型进行修正加项的方式,求取精确数学模型,并给出了其应用.

M odelD esign ofA ccurate M athem aticalR egression B ased on Experim entalD ata

In the production and scientific experiments, the relationship between the factors and indicators, i.e., it is difficult to accurately determine the mathematical model. For factors level division, through the experiments to seek its mathematical model, the mathematical model is the main form of linear regression equation, the experiment to determine the number of undetermined coefficients of mathematical model. The more factors and experiment, this method cause the larger error between analysis model and practical model, and even results in actual production application immeasurable loss. In this paper, the regression analysis model of increasing number of factors, the revised to calculate the precise mathematical model, and gives its application.