On Jul 11, 9:43=A0am, "AngleWyrm" <anglew...@[EMAIL PROTECTED]
> wrote:
> "Doug Wedel" <dougwe...@[EMAIL PROTECTED]
> wrote in message
>
> news:8NOdnXbn37Xx1fzVnZ2dnUVZ_u-dnZ2d@[EMAIL PROTECTED]
>
> > One definition of a paradox is "an apparently true statement or group
o=
f
> > statements that leads to a contradiction or a situation which defies
> > intuition"
> > Both Claude Shannon and Andrey Kolmogorov define the quantity of
> > information contained in a string in terms of the number of bits
requir=
ed
> > to specify the string. =A0This seems eminently reasonable -- the fewer
=
the
> > bits required to specify a string, the less "information" it contains
-=
- =A0
> > but it leads ineluctably to the conclusion that information is at its
> > maximum in random numbers.
>
> This re-casting of the definition of information leads to bad
conclusions=
..
> It is better to just use the word "data" when referring to bit strings.
>
> It is not reasonable to suggest that the length of a data string is
> pro****tional to the information it contains.
> The length of the string is not im****tant, since the measure is
normalize=
d,
> and the length is accounted.
> Lets have m objects. To their indexing by the binary decision tree
> we need at least log_2 digits. For 8 objects it is 24 digits,
> as 000, ... till 111.
> If objects are already indexed, e. g. aaaabbcd, the decision
> tree is shortened, and this shortening is information, we have about the
> string.
> To calculate the shortest decision trees were tedious, and thus it is
> more convenient to replace them by their limits, binary logarithms.
kunzmilan
|
