Re: information content of measures

From: Wei Dai <weidai.domain.name.hidden>
Date: Mon, 11 Jan 1999 01:35:39 -0800

Here is the fix I promised:

> C_U(v)=min {c:U'(u)=v and c=\sum u(p)(log u(p)+l(p))}
                ^
insert "u is a probability measure on infinite binary strings and". In
other words

u(\epsilon) = 1, and
u(x) = u(x0) + u(x1)
Received on Mon Jan 11 1999 - 01:38:16 PST