Question

我正在尝试对虚拟变量执行logit回归。

变量y是一个chr，我已使用以下代码将其转换为数字。

myData$y <- as.numeric(as.character(myData$y))
myData$x19 <- as.numeric(as.character(myData$x19))

结果如下：

> str(myData)
'data.frame':   2089 obs. of  21 variables:
 $ cod: chr  "DEFEB97411" "ESA87020566" "LT112021238" "ESA70302039" ...
 $ x1 : chr  "10.368400000000001" "18.104500000000002" "53.338142857142849" "18.096999999999998" ...
 $ x2 : chr  "81.251592378494607" "24.169624052591214" "20.258977190130423" "13.834562559615202" ...
 $ x3 : chr  "N/A" "N/A" "18.122857142857143" "15.467500000000001" ...
 $ x4 : chr  "92.257276111020005" "80.742324540987497" "53.251824749595997" "78.217022461160269" ...
 $ x5 : chr  "N/A" "N/A" "1.8639827074905999" "11.306989139206143" ...
 $ x6 : chr  "N/A" "N/A" "3.8017284097507149" "6.2387112552580009" ...
 $ x7 : chr  "N/A" "N/A" "-18.776333013159142" "63.587694201206432" ...
 $ x8 : chr  "N/A" "N/A" "N/A" "N/A" ...
 $ x9 : chr  "423.45434977976021" "11.683217643759502" "12.353727918269428" "7.4528876298177149" ...
 $ x10: chr  "188.66242539565681" "11.282744548351999" "11.902544271951285" "6.8315087254777138" ...
 $ x11: chr  "34.814814814814667" "0.71972866025899995" "1.4488644105372859" "1.128749296162143" ...
 $ x12: chr  "29.512676056338002" "11.225379651791" "10.349080357197428" "-48.646148547886426" ...
 $ x13: chr  "0.31797384757359998" "70.477571651141247" "81.34607558631528" "65.442536245008867" ...
 $ x14: chr  "347.83320016228879" "33.855519321625252" "27.042740986499286" "6.5450441963748585" ...
 $ x15: chr  "64.731000000000009" "10.11375" "10.614714285714285" "6.2912857142857135" ...
 $ x16: chr  "2.5469999999999997" "5.9179427965575062" "11.861177222645995" "5.9179427965575062" ...
 $ x17: chr  "5" "5.8" "6.2" "5.8" ...
 $ x18: chr  "5.2" "6.6943242163984618" "10.173227449851671" "6.6943242163984618" ...
 $ x19: num  "72.317277605096791" "68.303293784830231" "72.485515585867958" "68.303293784830231" ...
 $ y  : num  0 0 0 1 0 0 0 0 0 0 ...

然后，我正在尝试进行概率回归：

m1=glm(y~x19, 
       family=binomial(link="logit"))

但是我仍然收到错误消息：

> m1=glm(y~x19, 
+        family=binomial(link="logit"))
Error in weights * y : non-numeric argument to binary operator

您能帮我解决这个问题吗？

谢谢

Answer 1

以下是将myData直接传递给glm的方法：

m1=glm(y~x19, family=binomial(link="logit"), data=myData)

Answer 2

请在this other thread中查看我的帖子

glm()仅接受numeric或factor类型的变量。

权重错误* y：二进制运算符的非数字参数

2 个答案: