On Tue, 11 May 2004 10:33:16 +0200, "Konrad Den Ende"
<[EMAIL PROTECTED]> wrote:
>>>>> L = \sum_{i=1}^b \sum_{j=1}^k (y_ij - \beta_i - \mu_j)^2
>>>
>>> The partial derivatives i get are:
>>> dL/d(beta_i) = -2 \sum_{i=1}^b \sum_{j=1}^k (y_ij - \beta_i - \mu_j)
>>
>> Here's your problem right here. This is not correct.
>
>I've been looking at the expression for a while now and i can't see where i
>make a mistake. Trusting that you see where i gone wrong, i'll review my
>steps.
>1) The 2 "jumps" down in front of the paranthesis (that's f').
>2) Since it's just a constant i let it stand in front of the sums.
>3) g is the contents of the parenthesis and everything there is a constant
>with respect to i except for beta. It's derivative is clearly -1 and, as in
>(2) i let it get in front of the sums.
>That gives me the answer as i presented. No idea where it's wrong...
You've used "i" with two different meanings. Take the derivative
w.r.t. beta_n instead.
Duncan Murdoch
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
