摘要
回归正交设计法是研究盐胀率预报模型的常用方法之一,而回归正交设计法的适用性以及回归方程的优缺点分析却少有涉及。以现有试验数据为基础,进一步做回归方程的方差分析和SPSS逐步回归分析,具体研究回归方程显著性低、显著性系数可靠度低、计算值与实测值误差大、回归系数波动性大等缺点以及减少试验量、计算简单、影响因素直观、体现交互作用等优点,得出二次回归方程可作为盐胀率预报模型的粗略估计。更准确的预报及规律研究,需要建立更加准确的计算公式。
Regression orthogonal design is one of the methods commonly used to study prediction model of salt expansion rate, but its applicability, as well as the advantages and disadvantages of regression equation, are seldom studied. Based on existing test data, both variance analysis and SPSS stepwise regression analysis of the regression equation were further carried out to make detailed studies on the disadvantages thereof, including low significance of regression equation, low reliability of significance coefficient, large error between calculated and measured values, large fluctuation of regression coefficient, and the advantages thereof, including decreased test quantity, easy calculation, directly perceived affecting factors, embodied interaction. The quadratic regression equation obtained in such way may be taken as the rough estimate of the prediction model of salt expansion rate, and the more accurate prediction and rules study needs establishing more precise calculation formula.
出处
《路基工程》
2015年第4期44-48,共5页
Subgrade Engineering
关键词
盐渍土
盐胀率
回归正交设计法
方差分析
SPSS逐步回归分析
saline soil
salt expansion rate
regression orthogonal design
variance analysis
SPSS stepwise regression analysis