Say, A and B are two normally distributed parameters with their variations being $\sigma^2_a$ and $\sigma^2_b$. Now for system C, which is linearly dependent on these parameters, is its $\sigma^2_c=\sigma^2_a+\sigma^2_b$, or $\sigma_c=\sigma_a+\sigma_b$. To me adding parameter deviations seems natural. But what is the actual behaviour (and why)?
1 Answer
$\begingroup$
$\endgroup$
1
If A and B are normally distributed random variables that are independent and have variances σ$^2$a and σ$^2$b respectively and both have mean 0 then if C=A+B, C will be normally distributed with mean 0 and variance σ$^2$a+σ$^2$b. So σc=√(σ$^2$a+σ$^2$b) and not σa + σb. The reason is that for independent random variables the variance of the sum is the sum of the variances.
-
$\begingroup$ thanks. Also the discussions Intuitive explanation of variance and moment in Probability and Usefulness of Variancereally helped in clarifying the doubts I had about deviation and variance. Thanks again $\endgroup$– manuJun 8, 2012 at 20:22