×

On the network size for neuro-based stock option forecasting system: Experimental evaluation. (English) Zbl 0878.90015

Christer, Anthony H. (ed.) et al., Stochastic modelling in innovative manufacturing. Selected papers of the UK-Japanese workshop, Cambridge, UK, July 20–21, 1995. Berlin: Springer. Lect. Notes Eng. 445, 337-347 (1997).
Summary: This paper treats a forecasting system of European stock option prices based on the multi-layer neural network. In particular, the dependence of the number of input data on the system performance is examined. Finally, using the actual option price data traded in the Japanese market, the suitable network size, which consists of the numbers of the input data and the cells in the hidden layer, is determined in numerical experiments. We show that the neuro-based forecasting systems with some specified sizes are superior on their prediction abilities to the well-known Black-Scholes option formula.
For the entire collection see [Zbl 0855.00025].

MSC:

91B84 Economic time series analysis
91B24 Microeconomic theory (price theory and economic markets)
PDFBibTeX XMLCite