Re:Why squared error?
by cafebeen on Saturday May 16, 2015 @06:04PM
(#9557817)
My point was that many things are normally distributed (likely due to the CLT) and the square error is the right choice in those cases. This is an empirical thing as well, in that measurement noise is often Gaussian.
[ Parent
| on HN ]
There are no comments on this comment yet.