Re: HW3
4 views
Skip to first unread message
Percy Liang
Oct 11, 2007, 5:42:55 PM10/11/07
to Bill Burke, cs281a...@googlegroups.com
In the problem, we defined X = (X_1, \dots, X_n) as all the nodes.
The convention is that uppercase letters denote random variables and
lowercase letters denote the corresponding values.
Therefore, in this case, x corresponds to the value that X can take
on, which is a list of $n$ integers.
To be complete, we would write P(X = x), but instead, we abbreviate
it with p(x).
The convention is that uppercase letters denote random variables and
lowercase letters denote the corresponding values.
Therefore, in this case, x corresponds to the value that X can take
on, which is a list of $n$ integers.
To be complete, we would write P(X = x), but instead, we abbreviate
it with p(x).
\theta is all the parameters.
-Percy
On Oct 11, 2007, at 2:32 PM, Bill Burke wrote:
> Percy -- In problem 4, the model is referred to as p(x | theta).
> What does the lower case x mean? It is all of the nodes, a single
> node, or something else? Also, I am assuming that 'theta' without
> any subscripts means all of the theta's. Is this correct?
>
> Thanks,
> bill burke
Percy Liang
Oct 13, 2007, 1:10:16 AM10/13/07
to Xianzhi, cs281a...@googlegroups.com
> For Question 1.3, do we need to derive the derivative of the gamma
> function?
> function?
Put your answer in terms of the digamma function, which is the
derivative of the log-gamma function.
http://en.wikipedia.org/wiki/Digamma_function
-Percy
Reply all
Reply to author
Forward
0 new messages