I am trying so calculate the mean integrated squared error of my non-parametric estimated density function. I am estimating a exponentially distributed function.
For that i need to do the integral of $E((f_p(x, h) - f(x))^2)$
I know how to get $f_p(x,h)$ (gaussian kernel estimation) but i have no idea how to get f(x) for the density curve. I found how to get p(x) but i don't think it is what i need.
Nevermind i confused myself but in fact it's pretty easy :
$f(x) = \exp(-x)$ (if $\lambda=1$)