R - 如何按组进行回归并获得预测值?

时间:2015-02-27 02:39:05

标签: r statistics

personID<-c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26)
genger<-c('male', 'male', 'male', 'male', 'male', 'male', 'male', 'male', 'male', 'male', 'male', 'female', 'female', 'female', 'female', 'female', 'female', 'female', 'female', 'female', 'female', 'female', 'female', 'female', 'female', 'female')
height<-c(181, 161, 198, 195, 177, 175, 197, 195, 198, 193, 161, 167, 132, 181, 165, 151, 163, 180, 169, 181, 177, 135, 143, 107, 161, 142)
weight<-c(165,  73, 90, 89, 80, 159,    179,    177,    180,    175,    73, 76, 60, 165,    150,    69, 148,    164,    154,    165,    161,    61, 130,    97, 146,    65)
data<-data.frame(personID, genger, height, weight)
data

我是R初学者。

我喜欢按性别(男性,女性)执行回归。

回归公式为weight = solpe * height + intercept。

我做谷歌搜索,但我不理解几篇文章。

我想要的输出如下所示。

person_id   gender  height  weight  predict_value  error
1            male   181      165       xxx           xx
2            male   161      73        ...           ...  
3            male   198      90 
4            male   195      89 
5            male   177      80 
6            male   175      159    
7            male   197      179    
8            male   195      177    
9            male   198      180    
10           male   193      175    
11           male   161       73    
12          female  167       76    
13          female  132       60    
14          female  181      165    
15          female  165      150    
16          female  151       69    

如何按性别进行回归分析并添加预测和错误列?

任何帮助都会得到满足。

1 个答案:

答案 0 :(得分:2)

这是一种方式。您可以拆分数据,执行回归并使用predict()来查找置信区间,然后您可以将非拆分返回到原始结构。例如,使用您的测试数据并拆分样本数据中的“genger”(sic)列

unsplit(lapply(split(data, data$genger), function(x) {
    m<-lm(weight~height, x)
    cbind(x, predict(m, interval ="confidence"))
}), data$genger)

返回

   personID genger height weight       fit        lwr       upr
1         1   male    181    165 124.17126  94.106766 154.23576
2         2   male    161     73  87.11321  29.280886 144.94554
3         3   male    198     90 155.67061 115.126629 196.21458
4         4   male    195     89 150.11190 113.707198 186.51660
5         5   male    177     80 116.75965  83.508504 150.01080
# etc...