Let $X$ be a random variable with true density $f$, $Y = \{y_i\}_{i=1}^n$ be a realization of a random variable in $d$-dimensional space $R^d$, and $\hat{f}$ is the density estimator of $Y$ using kernel density estimation.
My question is what is the bound of $MISE(f,\hat{f})$ or $AMISE(f,\hat{f})$ with respect to optimal bandwidth matrix $H$ and dimensionality $d$?
