回複：回複：請問忍者豬和忍者菜及各位忍者X個專業問題：

- input data sampling is normalized(GAUSSION) (which is very simple one)
- calculate response thru software tool. what you get is a distribution of the response, in general(not always), a normal distribution as well
- data sampling normally is built-in the tool, however, you might generate the normalized data yourself (you have total control what kind of input data sampling) thru external tool, DOE appraoch
- once you got your response, you can do RSM (response surface methodology).
- a much better (than MONTE CARLO) data sampling is LHS (latin hyper cube)
hope it helps