Visitor's Comments

Visitors' Comments on Tutorial: LearningHistogram

We have 2 comments on this tutorial
Question > A question on the 'learning' example (date: 2012-01-06)
By Jialu Pu
in the Japanese character game, why we divide the reward by its number of frequency to compute the probability of a character, i.e h/tao/sigma(t/tao) Shouldn't h*tao/sigma(h*tao) makes more sense? If we regard reward as a weight.
rate up: 274 | rate down: 222 | last rated date: 2018-01-17
Is this comment useful? yes | no
> (date: )
rate up: 1 | rate down: 2 | last rated date: 2018-01-16
Is this comment useful? yes | no

These tutorial is copyrighted .