V. Gatón Bustillo, J. de Frutos
A financial derivative is a contract between two parts, whose price is dependent upon one or more underlying assets (stocks, bonds…). When we employ a complex model to replicate observed stock’s properties, we usually lack of an explicit formula for the derivative price, which has to be obtained by numerical methods. These methods usually require quite a big computational power in order to be competitive when applied to on-line trading. We develop a fast tensorial Chebyshev polynomial interpolation, where the storage problem (``Curse of Dimensionality'') is drastically reduced through a Reduced Bases Approximation method. The result will be a low storage cost polynomial which is able to price derivatives (or calibrate parameters) very fast for several stocks at the same time. The method is general enough to apply it with different models for stock dynamics or option types. We will test it for the NGARCH(1,1) model, which has 8 parameters, and use it to price S\&P500 European Options.
Keywords: Option Valuation, real-time trade, Reduced Bases
Scheduled
TE1 Computational Management Methods
May 31, 2016 4:45 PM
Salón de actos
