Is there any newer or better method (instead of Parzen) for estimating a probability density function from a set of data? for example can we use Copula?

So I assume by Parzen you mean Kernel Density estimation. You would have to actually provide some detail for what you mean by "better" method because there are approximately 8000 ways to assess "better". Are we talking about univariate data? Because if so I don't see how one could use copulas.

So I assume by Parzen you mean Kernel Density estimation. You would have to actually provide some detail for what you mean by "better" method because there are approximately 8000 ways to assess "better". Are we talking about univariate data? Because if so I don't see how one could use copulas.

Yes I mean Kernel
Actually I am working on a GRNN (general regression neural network) and I wonder if I can use a better pdf estimator (I mean more accurate results) rather than the kernel which is used by default.
Thank you