Apre？amento de op？？es sobre taxa de cambio R$/US$ negociadas no Brasil: uma compara？？o entre os modelos Black e redes neurais artificiais

in this study, a multilayer neural network model was applied to the pricing of r$/usd exchange rate call options traded on the s？o paulo securities, commodities and futures exchange (bm&fbovespa) from january 2004 to december 2007. based on the actual market prices, the performances of a neural network model and the black model were compared, using the usual error metrics and statistical tests. overall, the results showed that the artificial intelligence model outperformed the black model for the different degrees of moneyness.