[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Naive question
The short answer is "sort of." There are minimum description length measures
(how big is the
smallest automaton of a given class that generates a set of strings), but in
the general case while you can *define* the measure, you can't necessarily
effectively *compute* the measure. Caveat -- that's an off-the-cuff answer,
but if you have access to something like INSPEC you can search for "minimum
description length" and find more technical detail.
From: Bruce Grant [mailto:bgrant@xxxxxxxxxxxxx]
Sent: Wednesday, January 16, 2002 8:19 PM
Subject: Naive question
Are there other measures of information content than entropy?
As I understand it, if you interpret "in" as two characters instead of one,
entropy of the message containing it ought to appear lower because there are
more possible combinations of characters (i.e. "ii" or "nn" is possible if
and "n" are separate characters, but not if "in" represents a single
Is there such a thing as a measure of information content which is not so
dependent on knowing exactly what is or is not a character?