Calculation for Standard Deviation Given a Gaussian Distribution

148 Views Asked by At

Quick background: I work as a Software Engineer in the Embedded Systems space, and my primary job function has me working with signal processing algorithms. I feel pretty good when it comes to statistics, but this one comment left in a codebase by a former employee has me stumped.

The calculation is supposed to be for the standard deviation of a signal to noise difference, and is described as such: sigma = pow(1-normdist(n), m).

I haven't been able to place the equation to anything standard deviation related, at least from my knowledge. The 1-normdist(n) part remind me of calculating tail probabilities, but even then, taking a power of a probability doesn't make much sense to me.

Any help and advice is welcomed!

1

There are 1 best solutions below

0
On

Building on Paul's comment about $(1-p)^m$

and assuming normidst is the standard normal CDF $\Phi$ in some programming language

you have $(1-\Phi(n))^m$ as the probability that all $m$ independent samples from a standard normal distribution are at least $n$