r - 预测多个类别的值

Question

我使用下面的代码对两组进行线性回归，以根据年龄预测收入

asd <- data.frame(Age = c(20, 40, 50, 90, 10), Income = c(12, 24, 23, 26, 23), group = c("A","A","A","B","B"))
model <- asd %>% group_by(group) %>% do(model = lm(Income ~ Age, data = .))

但是在上一步之后，我需要计算 2 组的新预测收入

预期产出

  Age Income group    Predicted Income
1  20     12     A        XX            ##based on group A coefficients 
2  40     24     A        YY            ##based on group A coefficients 
3  50     23     A        DD            ##based on group A coefficients 
4  90     26     B        PP            ##based on group B coefficients 
5  10     23     B        RR            ##based on group B coefficients 

标签： r

Answer 1

像这样的东西？

asd %>% group_by(group) %>% 
  mutate(predicted_income = predict(lm(Income ~ Age)))

# A tibble: 5 x 4
# Groups:   group [2]
    Age Income group predicted_income
  <dbl>  <dbl> <chr>            <dbl>
1    20     12 A                 13.0
2    40     24 A                 21  
3    50     23 A                 25  
4    90     26 B                 26.0
5    10     23 B                 23.0