我正在如下构建Logistic回归(在R中):
glm_hr1 = glm(attrition_value ~ BusinessTravel+Department+Gender,
binomial(link="logit"), data=hr1)
X变量的定义如下:
BusinessTravel
--------------
[type=Factor]
[values="Frequently","None","Rarely"]
[Ref group = "Frequently"]
Department
----------
[type=Factor]
[values="HR","RD","Sales"
[Ref group = "HR"]
Gender
------
[type=Factor]
[values="Male","Female"]
[Ref group = "Male"]
summary(glm_hr1)返回以下内容:
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) -0.8168 0.2476 -3.298 0.000973 ***
BusinessTravelNone -1.3738 0.2354 -5.836 5.35e-09 ***
BusinessTravelRarely -0.6464 0.1166 -5.545 2.94e-08 ***
DepartmentRD -0.3805 0.2387 -1.594 0.110906
DepartmentSales 0.1187 0.2441 0.486 0.626912
GenderFemale -0.2086 0.1052 -1.982 0.047496 *
如何获得的系数: 经常商务旅行 人力资源部 性别男
全部汇总?
我使用-1来建立模型,如下所示:
glm_hr2 = glm(attrition_value ~ BusinessTravel+Department+Gender -1,
binomial(link="logit"), data=hr1)
此模型的输出为:
Coefficients:
Estimate Std. Error z value Pr(>|z|)
BusinessTravelFrequently -0.8168 0.2476 -3.298 0.000973 ***
BusinessTravelNone -2.1905 0.3101 -7.063 1.63e-12 ***
BusinessTravelRarely -1.4632 0.2333 -6.272 3.56e-10 ***
DepartmentRD -0.3805 0.2387 -1.594 0.110906
DepartmentSales 0.1187 0.2441 0.486 0.626912
GenderFemale -0.2086 0.1052 -1.982 0.047496 *
(拦截)一词消失了,由“ BusinessTravelFrequently”代替。 我知道为什么会这样。
更改公式的顺序,但是每个级别的游戏系数 例如:
glm(attrition_value ~ Department+BusinessTravel+Gender -1, ....)
为部门的所有值赋予系数
glm(attrition_value ~ Gender+Department+BusinessTravel -1, ....)
为所有性别值赋予系数
有没有办法在一个摘要中获得所有因子变量的所有系数?
还有其他方法吗?