Question

我使用线性回归建模我的数据。我想多次运行Bonferroni离群值测试并从我的数据中删除相应的记录。我的问题是：我无法从outlierResult中提取id。这是可重现的代码。我想根据伪代码编写一个while循环。我在R中编码。

# URL <- "http://www.math.uah.edu/stat/data/Galton.csv"
# download.file(URL, destfile = "./galton.csv", method="curl")
galton <-read.csv("galton.csv")
attach(galton)

dim(galton)
head(galton)

##creating outliers
set.seed(1)
random_index <- sample(1:nrow(galton), size = 5, replace = FALSE, prob = NULL)
print(random_index)
galton[random_index,"Height"] = galton[random_index,"Height"] +100

set.seed(2)
random_index2 <- sample(1:nrow(galton), size = 5, replace = FALSE, prob = NULL)
galton[random_index2,"Height"] = galton[random_index2,"Height"] +75

set.seed(3)
random_index3 <- sample(1:nrow(galton), size = 5, replace = FALSE, prob = NULL)
galton[random_index3,"Height"] = galton[random_index3,"Height"] +50


linear_reg <- lm(Height~Father+Mother,data=galton)

require(car, quietly=TRUE)
outlierResult <-outlierTest(linear_reg)
outlierResult


# the pseudocode
# while outlierResult is not empty
#   remove the corresponding records
#   linear_reg <- lm(Height~Father+Mother,data=galton)
#   outlierResult <-outlierTest(linear_reg)

Answer 1

请参阅下文。诀窍是要注意，如果我正确读取它，异常结果会给出行名称。

library(car, quietly=TRUE)
galton <-read.csv("http://www.math.uah.edu/stat/data/Galton.csv")
attach(galton)

dim(galton)
head(galton)

##creating outliers
set.seed(1)
random_index <- sample(1:nrow(galton), size = 5, replace = FALSE, prob = NULL)
print(random_index)
galton[random_index,"Height"] = galton[random_index,"Height"] +100

set.seed(2)
random_index2 <- sample(1:nrow(galton), size = 5, replace = FALSE, prob = NULL)
galton[random_index2,"Height"] = galton[random_index2,"Height"] +75

set.seed(3)
random_index3 <- sample(1:nrow(galton), size = 5, replace = FALSE, prob = NULL)
galton[random_index3,"Height"] = galton[random_index3,"Height"] +50




currentData <- galton
linear_reg <- lm(Height~Father+Mother,data=currentData)

outlierResult <-outlierTest(linear_reg)
outlierResult

while(length(outlierResult)!=0){
  exclusionRows <-names(outlierResult[[1]])
  inclusionRows <- !(rownames(currentData) %in% exclusionRows)
  currentData <- currentData[inclusionRows,]

  linear_reg <- lm(Height~Father+Mother,data=currentData)
  outlierResult <-outlierTest(linear_reg)

}

在循环中删除Bonferroni异常值测试结果

1 个答案: