反规范化 R 中的数据

Question

I apply normalization function to my data with this code:

normalize <- function(x) { return ((x - min(x)) / (max(x) - min(x)))}

Then I applied the function to my data, only for numeric variables.

bankfull[,c(1,6,10,12:15)]= as.data.frame(lapply(bankfull[,c(1,6,10,12:15)], normalize))

Can someone help me to revert normalization?

Thank you in advance.

Answer 1

There's a package DMwR that unscales sets of values and is worth a shot, even though there are pitfalls to watch out for like John mentioned in the comments. For example, with 100 observations normally distributed with mean of 5 and stdev of 1:

library(DMwR)
a <- rnorm(100,5,1)
hist(a)

b <- scale(a)
c <- DMwR::unscale(b,b)

Now you can compare the values by subtracting ac (you get all nearly-zeroes) or plotting a vs c (which shows the equivalents):