On Jul 24, 2009, at 11:23 AM, Isaac Deutsch <i...@gmx.ch> wrote:
An overall drift in the numbers might be nothing. Some pattern (sub)
sets can be multiplied by a constant value without affecting
overall prediction accuracy. Fixing one or more gamma values may
fix your drift issue. I think RemÃ's paper forced the average gamm
a to be 1 after each iteration.
If I fixed the average gamma, wouldn't that make the other features
drift *upwards*? I know it isn't a problem in general, but if the
number range is too big I might get precision issues.
Do you suggest I fix the value of a single 3x3 pattern? Would that
stabilize the 3x3 pattern values or would it also affect the other
values (extend, kill, etc.)?
To answer exactly, I need to know more about how you set up your
patterns. If every point gets one, and exactly one 3x3 pattern, then
fixing one 3x3 pattern is required. If some points have no 3x3
pattern, then you're implicitly fixing an "other" pattern to a value
of 1 and then no fixing of a 3x3 pattern is needed.
Just like 3x3 patterns, any orthogonal subset should have one value
fixed...
Does that help?
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/