Re: Entropy

[Leigh Palmer <palmer@SFU.CA>]

> I've been interested in the SETI project for a while and actually
> 'ran' some of the data analysis on my desktop computer.
>
> Does anybody have any clue as to how the program identifies
> 'potential hits' as distinct from the 'random noise'?
>
> We gotta be talking about Shannon entropy rather than 'micro-states'
> or else all signals would have equal probability and the whole SETI
> notion would be bogus. (perhaps it IS bogus?)

The Serendip receiver at Arecibo looks for a signal on a carrier. The
carrier will make its presence known by being an underlying gaussian
signal as a function of time as the sky scans across the receiver.
There is no way to separate nonrandom bits from random bits without
such a carrier signal being present. This is what seti@home is doing
in analyzing data from Serendip.

Leigh