Problem Dependent parameter passing into slicesample
2 views (last 30 days)
Is there a way for me to pass a problem dependent parameter into slicesample similar to how it is done for lsqnonlin, i.e.
x = lsqnonlin(@(x) myfun(x,c),[1;1]);
function F = myfun(x,c)
F = [ 2*x(1) - exp(c*x(1))
-x(1) - exp(c*x(2))
x(1) - x(2) ];
%hoping to do something like this:
x = slicesample([1;1],N,'pdf',f(x,c),'thin',5,'burnin',1000);
So far it seems like slicesample can only be called with the parameters that it will explore the pdf space with.
Prudhvi Peddagoni on 12 Mar 2021
You can pass additional parameters such as c to the slicesample function like this:
x = slicesample([1;1],N,'pdf',@(x) myfun(x,c),'thin',5,'burnin',1000);
The reason it is throwing error is because there are 3 different functions in myfun. In lsqnonlin function, it returns x such that the sum of squares of these multiple input functions is minimum. But in the case of slicesample function, you need to give a single function.
Hope this helps.