Dr Zheng Gong
National University of Singapore
We propose a nonparametric way of estimating of option prices and state price densities via wavelet, while still preserving the monotonicity and convexity of the option prices with respect to the strike price. A significant advantage of the current method over the local polynomial regression is that it can fit options with different maturities and strike prices simultaneously, by using one set of parameters instead of different sets of parameters for different maturities. Furthermore, due to local adaptive features of wavelet, numerically the proposed method has better out of sample prediction. An asymptotic analysis is also given.