The difference between Gaussian Noise variance (user defined) and MATLAB generated variance (using built-in function var)

I added Gaussian Noise (mean 0, variance 0.01) to a matrix of all zeros. The variance from using built-in function var is expected to be 0.01 as the noise addition. However, it is 0.0034. What is the difference / connection between these 2 variance values?

Answer 1

[Dimitris Kalogiros]

Dear Ann

Maybe you have using too litle data. For example , if you execute command var(randn(1,1000)) many consecutive time you will get different results that, also, differ from 1. But if you try var(randn(1,1E7)) , results tends to unity.

More over var(), has a flexible definition in matlab . Look at the following figure:

where "w" is a parameter of var() calling.

For more info , have a look at matlab help