On Mon, Mar 31, 2008 at 2:10 PM, Record, James N
<james.n.record at lmco.com> wrote:
> I am new to using maxima. I am trying to take teh partial derivative of this
> fuction:
>
> e(x,y,ybar,m,s):=sum(ybar[l]*product(exp(-1/2*(x-m[k])^2/s[k]^2),k,1,p),l,1,M)/sum(product(exp(-1/2*(x-m[k])^2/s[k]^2),k,1,p),l,1,M)-y;
>
> w.r.t m (for example).
Jim, the product has no dependence on l so it factors out of the
summations and cancels. So e(x,y,ybar,m,s) reduces to
sum(ybar[l], l, 1, M)/M - y which has no dependence on m.
Not sure what you intended there.
Robert Dodier