A strategy for forecasting option prices using fuzzy time series and least square support vector regression with a bootstrap model

Recently, the strategy for forecasting option price has become a popular financial topic because options are important tools on risk management in financial investments. the well-known black-scholes model (b-s model) is widely used for option pricing. in b-s model, the normal distribution assumption is important. however, the financial data in real life may not follow the normal distribution assumption. for this reason, this paper proposes a novel hybrid model, which is a nonlinear prediction model without normal distribution assumptions to forecast the option prices. the proposed model is composed of a fuzzy time series (fts) model, a least square support vector regression (lssvr), and a bootstrap method. in the proposed model, fts model is firstly used to fuzzily dataset and to build historical database. subsequently, the proposed method uses the remainder contractual time to search similar historical trends as training samples. finally, we use the bootstrap method on lssvr to enhance the prediction accuracy. the experiment results show that the proposed model outperforms traditional time series models and several hybrid models in terms of the root mean square error (rmse), the mean absolute error (mae) and the correlation coefficient (r) of actual and forecasted option price.