SEMINARIESCHEMA FÖR MATEMATISK STATISTIK
Torsdag 28/3 14.15 Henrik Amilon,
1996 Mat stat i Lund, presenterar sitt examensarbete:
Optionsvärdering med Neurala Nätverk
Abstract:
The Black-Scholes formula is a well-known valuation model for pricing
and hedging (european) options. It relies, however, upon several
assumptions which are highly questionable. The thesis examines whether
an Artificial Neural Network (ANN) could be used to find a pricing
formula with better correspondence to market data than the traditional
Black-Scholes formula. As a first step the investigation concerns the
possibility for a Multi Layer Perceptron (MLP) to recover the
Black-Scholes formula for call options by training it on simulated
Black-Scholes option prices. The results from this attempt are
fruitfull. At a second stage the network method is applied to pricing
and delta-hedging of swedish OMX call options from 1990-1994. As a
benchmark estimates generated by Black-Scholes model are used.
Comparisons reveal that MLP:s do not consistently give a more accurate pricing, or
a better hedge, than the Black-Scholes formula.
Lokal: Rum 313 i Mattehuset.
Björn Holmquist 046-2228546