my question is exactly as follows
library(caret)
data(cars)
head(cars)
colnames(cars)
Expected Answer
cars$type <- names(cars[14:18])[max.col(cars[14:18])]
but using the name of the columns such as below which is not working. how to get around with this? many thanks in advance.
cars$type <- cars[c("convertible", "coupe", "hatchback", "sedan", "wagon" )][apply(cars[c("convertible", "coupe", "hatchback", "sedan", "wagon" )], 1, match, x = 1)]
head(cars)
You can subset dataframe by name in the similar fashion -
cols <- c("convertible", "coupe", "hatchback", "sedan", "wagon" )
cars$type <- cols[max.col(cars[cols])]
To check the output -
identical(cols[max.col(cars[cols])] , names(cars[14:18])[max.col(cars[14:18])])
#[1] TRUE
It is better to specify the ties.method or else it can choose "random" as default and this could change the outcome in each run where there are multiple max values per row
cols <- c("convertible", "coupe", "hatchback", "sedan", "wagon" )
cars$type <- cols[max.col(cars[cols], "first")]