# generate Gaussian noise with certain mean and variance

156 views (last 30 days)
Na on 19 Jul 2019
Commented: Yash Totla on 19 Jul 2019
I want to produse Gaussian noise. So, I use normrnd.
m=0.17;
n=normrnd(6.*m,m.*m);
I think the n should be sometimes negative number as it is Gaussian. But always be positive number. So normrnd is correct way to do this?

Yash Totla on 19 Jul 2019
The mean of your Gaussian distribution is 1.02 and the standard deviation is 0.17.
This means that 68% area under the bell curve will be between 0.85 and 1.19 ( ), the region between the values 0.68 and 1.36 will make up for almost 96% of the area under the curve, and the 99% of the bell curve will lie between 0.51 and 1.53 .
This means that there is only 0.5% chance that you get a value less than 0.51 and the chance of getting a negative value will be negligible (you can integrate ~ from to 0 and find the probablity of getting a negative number).
Na on 19 Jul 2019
Thank you for the explanation. It mean that insted of normrnd(6.*m,m.*m), I use norm(6.*m,m.*m).
Yash Totla on 19 Jul 2019
From what I know, Gaussian white noise has a mean at 0. Why do you want to add noise which has mean at 1.02? Can you elaborate more on this please? For what reason are you adding Gaussian noise? Can you provide more insight on the problem statement?