Learn How to Calculate SNR for Vector A in Time Series | Tips & Tricks

[sue132]

Hi,

I have a vector 'A' of length 'n', whose values change with time. 'A' gives the change from a baseline 'B'. I calculate a signal as the distance A from the average of B, i.e.,

If 'B' is my baseline, then the signal at a time 't' due to A(t) would be given by S=(1/n)*(Ʃ₁A_i -<B> - Ʃ₂A_i -<B>), i=1 to n. 'A' is divided into two groups of elements whose values are higher than their previous values and those whose values are lower.

I define noise as the standard deviation of S.

Given this, how do I define the signal-to-noise ratio of A?

I found different ways of calculating the SNR in different contexts, but am not very clear as to which one to use in this context.

Thanks in advance for any suggestions/help.

 

[mfb]

Dividing S by the standard deviation of S should give a meaningful quantity, if the formula for S itself is meaningful (it has missing brackets here, by the way).