SPD and RND estimation

As a cooking recipe, will probably benefit a lot from first converting 

> the prices into Black-Scholes implied volatilities and then smoothing
 > the implied volatilities, after which you can go back into price space
 > (don't worry about unknown dividend yield - you can back it out from
 > put-call parity).

The problem with this approach is that the point of the cited 
Ait-Sahalia and Lo paper is to do a non-parametric estimation of the 
density function from prices. They propose the use of kernel smoothing 
regression, although other non-parametric methods are cited in other papers.

Regards,

   - Brian

Ref:
Ait-Sahalia, Yacine and Lo, Andrew W., "Nonparametric Estimation of
State-Price Densities Implicit in Financial Asset Prices" (November
1995). NBER Working Paper No. W5351.
Available at SSRN: http://ssrn.com/abstract=225414

-----Original Message-----
From: Sankalp Upadhyay
Sent: Sunday, July 15, 2007 8:55 AM

Hi,

I am trying to do an estimation of State Price Density (SPD) and Risk
Neutral Density (RND) from a set of option prices - preferably in a
non-parametric way.
Is there some package in R that can help? fOptions does not seem to be have
this.

Alternatively, would you know a standard method or a very good research
paper/article/reference on this topic? The utility being that a good
paper/article can be changed to R code easily.

Many thanks,

Sankalp