如何将列表传递给R中的函数

Question

我想将线性回归的结果存储在截距和系数等列表中。我有以下代码

results  = list()
intercept = list()
coeff = list()

GetLM = function(dataframe,results,intercept,coeff){
     unq_clients = as.vector(unique(dataframe$clients))
     for(i in 1:length(unq_clients)){
         new_df=dataframe[dataframe$clients ==    unq_clients[i],]
         regression= lm(a ~ b,data = new_df )
         results[[i]] = coef(regression)
         intercept[i] = results[[i]][1]
         coeff[i]=results[[i]][2]
     }
}

但是当我调用该函数时，列表中没有存储任何内容。我做错了什么？

Answer 1

我觉得你以非 R 方式编写了这个函数。为什么需要results、intercept 和coeff 作为GetLM 的参数？相反，让函数返回一个包含这些变量的列表：

GetLM <- function(dataframe){
     unq_clients <- as.vector(unique(dataframe$clients))
     for(i in 1:length(unq_clients)){
         new_df <- dataframe[dataframe$clients == unq_clients[i],]
         regression <- lm(a~b,data = new_df )
         results <- coef(regression)
         intercept[i] <- results[[i]][1]
         coeff[i] <- results[[i]][2]
     }
     return(list(results = results, intercept = intercept, coeff = coeff))
}

所以现在你可以这样做了：

reslist <- GetLM(dataframe = your_data_frame)
results <- reslist$results
intercept <- reslist$intercept
coeff <- reslist$coeff

Answer 2

您可以完美地继续使用您自己的代码，只需添加我在 cmets 中提到的 return() 语句。

这里我使用lapply() 替换for 循环，并使其更通用，即使您的模型有多个独立的，您也可以以更好的方式提取它们。

注意：我使用mtcars 解释，因为您没有与我们分享任何数据

unq_cyl = as.vector(unique(mtcars$cyl)) # 3 unique values
results <- lapply(seq_along(unq_cyl), function(x) {
                                    new_df = mtcars[mtcars$cyl == unq_cyl[x],]
                                    regression= lm(mpg ~ disp + wt,data = new_df )
                                    results=coef(regression)
                                    results})

intercept <- sapply(results, FUN = function(i) i[names(i) %in% "(Intercept)"])
coeff <- sapply(results, FUN = function(i) i[!(names(i) %in% "(Intercept)")])


results
#[[1]]
#(Intercept)        disp          wt 
#28.18834771  0.01914227 -3.83509580 
#
#[[2]]
#(Intercept)        disp          wt 
# 41.1350198  -0.1224733  -0.6978035 
#
#[[3]]
# (Intercept)         disp           wt 
#24.077855175 -0.002511735 -2.023137804 

intercept
#(Intercept) (Intercept) (Intercept) 
#   28.18835    41.13502    24.07786 
coeff
#            [,1]       [,2]         [,3]
#disp  0.01914227 -0.1224733 -0.002511735
#wt   -3.83509580 -0.6978035 -2.023137804