[R] kernel density estimation and classification

[wk y]

Hi,
 
I am interested to use Kernel Density Estimation on bivariate data for multi-class classification.
 
So far, I have managed to use the 'ks' package to plot the contours of the kernel density estimates based on 8-class training dataset with only 2 variables. 
 
However, I do not know how to make use of the Kernel Density Analysis results as input into a classifier (e.g. Bayesian classifiers) to accomplish the task of classifying testing dataset into 8 different classes.
 
I will really appreciate any suggestions or advice.
 
Thanks.
 
regards,
WK