如何在R中使用多元回归时仅返回单个预测值

Question

我遇到一个问题，在尝试预测R中的单个结果时会得到数百个结果。有关如何解决此问题的任何建议都将非常有帮助。

logwage <- log(wage_data$wage, 10)

mlogwage <- lm(logwage ~ occupation + education + experience + age + south,
               data = wage_data)

newlogdata <- data.frame(occupation= "Sales", 
                         education = 16, 
                         experience = 10,
                         age = 45, 
                         south = "Yes", 
                         data = logwage)

predict(mlogwage,
        data = newlogdata, 
        interval = "confidence")

Answer 1

没有可复制的示例，很难重现该错误。因此，这是一个示例，说明如何使用mtcars数据集预测单个观测值。

aModel <- lm(mpg ~ am + disp + wt,data = mtcars)
aCar <- data.frame(am = 1,disp = 288,wt = 3.21)
predict(aModel,aCar,interval = "confidence")

...和输出：

> predict(aModel,aCar,interval = "confidence")
       fit      lwr      upr
1 19.20009 16.88843 21.51175
>