Reputation: 79
The iris
dataset (built-in in R) includes 50 observations, each observation has data of Sepal.Length, Sepal.Width, Petal.Length and Petal.Width. I want to use multiple tests (Shapiro.test, ks.test, cvm.test, ad.test, ) to test the normality of each column and show the result in on table.
The code below shows the result using only one test.
dat <- iris %>%
filter(Species == "setosa")
df <- dat %>%
select(-Species)
test <- lapply(df, shapiro.test)
table <- sapply(test, `[`,c("statistic","p.value"))
table
R >
$Sepal.Length.p.value
[1] 0.4595132
$Sepal.Width.p.value
[1] 0.2715264
$Petal.Length.p.value
[1] 0.05481147
$Petal.Width.p.value
[1] 8.658573e-07
I wish to summarize and compare between different tests in one table, where the row represents different test and the column represents Sepal.Length, Sepal.Width, Petal.Length and Petal.Width.
Upvotes: 2
Views: 472
Reputation: 887118
One option would be
library(nortest)
out <- lapply(c("shapiro.test", "cvm.test", "ad.test"),
function(x) sapply(df, function(y) get(x)(y)[c("statistic", "p.value")]))
do.call(rbind, Map(cbind, test = c("shapiro.test", "cvm.test", "ad.test"), out))
Or using tidyverse
library(tidyverse)
lst(shapiro.test, cvm.test, ad.test) %>%
map_df(~ df %>%
summarise_all(list(~ list(.x(.)[c("statistic", "p.value")]))) %>%
unnest, .id = "test")
If we don't need the intercept
lst(shapiro.test, cvm.test, ad.test) %>%
map_df(~ df %>%
summarise_all(list(~ list(.x(.)[c("statistic", "p.value")][2]))) %>%
unnest, .id = "test")
# test Sepal.Length Sepal.Width Petal.Length Petal.Width
#1 shapiro.test 0.4595132 0.2715264 0.05481147 8.658573e-07
#2 cvm.test 0.2596871 0.2324437 0.006874393 1.717773e-09
#3 ad.test 0.3352439 0.2101787 0.01079067 7.437223e-12
Upvotes: 0