How to computate the gini index by column when NA is there.

[Comede_way]

Hi,
how to compute the gini index use Desctools or other packages in R? I could only complete one column that not have na. Take the “col1”,“col2”, “col3” for example. Thank you in advance.

c1<-c(16.6,1.0,10.1,8.6,8.0,17.0,2.4,7.6,5.7,11.6,3.6,2.8,6.3,1.5,2.7,7.4,6.2,2.5,6.0,5.5,2.0,6.7,5.6,3.5,10.4,6.8,3.5,NA,NA)
c2<-c(9.3,10.6,1.3,2.8,5.8,NA,9.4,2.5,3.2,7.3,12.5,18.4,15.0,11.8,13.7,10.2,9.5,7.6,2.5,13.6,4.8,3.6,3.2,10.2,17.7,6.8,3.5,18.9,NA)
c3<-c(5.6,5.8,13.3,13.8,3.0,10.1,10.7,7.7,12.3,15.8,3.5,12.7,4.2,8.1,13.6,19.8,14.3,16.8,9.9,6.6,8.9,10.8,13.1,12.1,5.9,10.1,11.2,18.8,11.0)
c4<-c(1:29)
c5<-data.frame(c4,c1,c3,c2)
colnames(c5)<-c("id","col1","col2","col3")

^{Created on 2022-10-31 with reprex v2.0.2}

[technocrat]

The function signature, from help(Gini)

Gini(x, weights = NULL, unbiased = TRUE,
     conf.level = NA, R = 10000, type = "bca", na.rm = FALSE)

so, set

na.rm = TRUE

[Comede_way]

OK, as we could see, there are 3 columns in c5 dataframe. Is it possible to calculate gini index for each column using loop function, apply function or others? when we calculate gini index for col1, ignoring the NA. Others are the same. Thanks a lot.

@Comede_way

c1<-c(16.6,1.0,10.1,8.6,8.0,17.0,2.4,7.6,5.7,11.6,3.6,2.8,6.3,1.5,2.7,7.4,6.2,2.5,6.0,5.5,2.0,6.7,5.6,3.5,10.4,6.8,3.5,NA,NA)

[technocrat]

c1 <- c(16.6, 1.0, 10.1, 8.6, 8.0, 17.0, 2.4, 7.6, 5.7, 11.6, 3.6, 2.8, 6.3, 1.5, 2.7, 7.4, 6.2, 2.5, 6.0, 5.5, 2.0, 6.7, 5.6, 3.5, 10.4, 6.8, 3.5, NA, NA)
c2 <- c(9.3, 10.6, 1.3, 2.8, 5.8, NA, 9.4, 2.5, 3.2, 7.3, 12.5, 18.4, 15.0, 11.8, 13.7, 10.2, 9.5, 7.6, 2.5, 13.6, 4.8, 3.6, 3.2, 10.2, 17.7, 6.8, 3.5, 18.9, NA)
c3 <- c(5.6, 5.8, 13.3, 13.8, 3.0, 10.1, 10.7, 7.7, 12.3, 15.8, 3.5, 12.7, 4.2, 8.1, 13.6, 19.8, 14.3, 16.8, 9.9, 6.6, 8.9, 10.8, 13.1, 12.1, 5.9, 10.1, 11.2, 18.8, 11.0)

the_series <- list(c1,c2,c3)
my_gini <- function(x) DescTools::Gini(x, na.rm = TRUE)
the_ginis <- lapply(the_series,my_gini)
the_ginis |> unlist()
#> [1] 0.3537744 0.3464965 0.2344565

^{Created on 2022-11-08 by the reprex package (v2.0.1)}

[Comede_way]

It works!!! I am very grateful to your effort and time in this matter.