在 R 中具有名义和有序预测变量的拟合逻辑模型

Question

我正在尝试在 R 中拟合逻辑回归，我的回答是 学生是否处于留校察看期，我的六个自变量是：

x1 = gender(male,female) which is nominal.
x2 = Hour spend in family activity (0 hour, <5 hour, 5-9 hour, 10-14 hour, 
>14 hour)
x3 = number of times under probation
x4 = types of school(public, private, other)
x5 = level of attendance in class( 70%, 70-79%, 80-89%, 90-99%, 100%)
x6 = hours of library work (0 hour, <5 hour, 5-9 hour, 10-14 hour, >14 hour)

我的问题是我混合了名义和序数，有人可以帮助如何在 R 中适应这个模型吗？

Answer 1

只需将变量转换为所需的每种类型，对于这种情况：

x1 = factor(x1, levels = c("male", "female")) #nominal
x2 = factor(x2, levels = c("0 hour", "<5 hour", "5-9 hour", "10-14 hour", 
">14 hour"), ordered = TRUE) #ordinal
x3 #stay_numeric
x4 = factor(x4, levels = c("public", "private", "other"))  #nominal
x5 = factor(x5, levels = c("70%", "70-79%", "80-89%", "90-99%", "100%"), 
ordered = TRUE) #ordinal
x6 = factor(x6, levels = c("0 hour", "<5 hour", "5-9 hour", "10-14 hour", ">14 hour"), ordered = TRUE) #ordinal

然后您可以使用glm库将这些变量拟合到模型中，它将自动扩展变量的水平或类别。