论文部分内容阅读
两组判别分析是一种多变量的统计方法。它经过一定的线性组合把多个变量变换为一个新变量(判别函数)。这个新变量能使两组组间的差异最大,而两组组内的差异最小,从而能使两组最好地区分开。这种方法可以在事先知道两组样品的条件下,解决未知样品的归属问题。因此,它在划分胀缩土与非胀缩土、新近堆积黄土与一般湿陷性黄土、近代沉积土与一般粘性土等方面都有较明显的效果。
Two sets of discriminant analysis is a multivariate statistical method. It transforms multiple variables into a new variable (discriminant function) after a certain linear combination. This new variable maximizes the difference between the two groups, while the difference within the two groups is minimal, allowing the two groups to be best separated. This method can solve the ownership problem of unknown samples under the condition that two sets of samples are known in advance. Therefore, it has obvious effects in dividing shrinkage soil and non-flattened soil, newly accumulated loess and general collapsible loess, modern sedimentary soil and general clay soil.