使用Data.Table包按组拟合模型

时间:2015-04-02 19:12:33

标签: r data.table

如何使用data.table语法按组安装多个模型?我希望我的输出是一个data.frame,每个" by group"每个模型适合一列。目前我可以使用dplyr包执行此操作,但无法在data.table中执行此操作。

# example data frame
df <- data.table(
   id = sample(c("id01", "id02", "id03"), N, TRUE),     
   v1 = sample(5, N, TRUE),                          
   v2 = sample(round(runif(100, max = 100), 4), N, TRUE) 
)

# equivalent code in dplyr
group_by(df, id) %>%
do( model1= lm(v1 ~v2, .),
    model2= lm(v2 ~v1, .)
  )

# attempt in data.table
df[, .(model1 = lm(v1~v2, .SD), model2 = lm(v2~v1, .SD) ), by = id ]

# Brodie G's solution
df[, .(model1 = list(lm(v1~v2, .SD)), model2 = list(lm(v2~v1, .SD))), by = id ]

1 个答案:

答案 0 :(得分:7)

尝试:

df[, .(model1 = list(lm(v1~v2, .SD)), model2 = list(lm(v2~v1, .SD))), by = id ]

或稍微更具惯用语:

formulas <- list(v1~v2, v2~v1)
df[, lapply(formulas, function(x) list(lm(x, data=.SD))), by=id]