Descale预测值

Question

Ali是一个包含4个目标变量和16个预测变量的矩阵。我对所有变量进行了规范化，然后使用pcaNNet来获取网络预测。现在我如何将预测值除去原始尺度？感谢

library(clusterSim)#doing standardization
library(caret)
normAli = data.Normalization(Ali,type="n4")
# or this code : normAli = apply(Ali, MARGIN = 2, FUN = function(Ali) (Ali - min(Ali))/diff(range(Ali)))

normxmat<- normAli[2:16] # predictor 
normymat<- normAli[17:20] # targets

fit<- pcaNNet(normxmat,normymat, size =10 , linout = TRUE, trace = TRUE)


pred<- predict(fit)