引言随着计算机的普及,多元回归分析方法受到人们越来越广泛的重视。这一方法应用于天气予报,地质勘探,服装设计标准的制定方面,都收到了很好的效果。我们把该分析方法运用于水泥质量予测,并对招远水泥35批数据,选取了九个因子,两个予报量进行了实际计算,与经常使用的一元线性回归及多对一线回归方法相比有明显的长处。我们编制了在IBM--PC机上运行的BASIC语言程序,可进行逐步回归计算及多对多双重筛选的逐步回归计 Introduction With the popularity of computers, multiple regression analysis methods have received more and more attention. This method has been applied to weather forecasting, geological exploration, and the design of clothing design standards. We applied this analysis method to the prediction of cement quality. We selected nine factors for Zhaoyuan Cement’s 35 batches of data, and two of them reported actual calculations, and used the one-dimensional linear regression and multiple-pair first-line regression methods. There are obvious advantages compared to it. We have compiled a BASIC language program running on an IBM-PC machine that can perform stepwise regression calculations and multiple-to-many dual-filtering stepwise regression