kernel_density: multivariate kernel density estimator

 usage:
       dens = kernel_density(eval_points, data, bandwidth)

 inputs:
       eval_points: PxK matrix of points at which to calculate the density
       data: NxK matrix of data points
       bandwidth: positive scalar, the smoothing parameter. The fit
               is more smooth as the bandwidth increases.
       kernel (optional): string. Name of the kernel function. Default is
               Gaussian kernel.
       prewhiten bool (optional): default false. If true, rotate data
               using Choleski decomposition of inverse of covariance,
               to approximate independence after the transformation, which
               makes a product kernel a reasonable choice.
       do_cv: bool (optional). default false. If true, calculate leave-1-out
                density for cross validation
       computenodes: int (optional, default 0).
               Number of compute nodes for parallel evaluation
       debug: bool (optional, default false). show results on compute nodes if doing
               a parallel run
 outputs:
       dens: Px1 vector: the fitted density value at each of the P evaluation points.

 References:
 Wand, M.P. and Jones, M.C. (1995), 'Kernel smoothing'.
 http://www.xplore-stat.de/ebooks/scripts/spm/html/spmhtmlframe73.html

Package: econometrics