Forecasting stock price based on fuzzy time-series with equal-frequency partitioning and fast Fourier transform algorithm

The prediction of stock markets is an important and widely research issue since it could be had significant benefits and impacts, and the fuzzy time-series models have been often utilized to be the forecast models to make reasonably accurate predictions. For promoting the forecasting performance of fuzzy time-series models, this paper proposed a new model, which incorporates the concept of the equal-frequency partitioning and fast Fourier transform algorithm, and it's never be adopt on fuzzy time-series before. In order to evaluate our proposed approach, the source data was using actual trading data from Taiwan Stock Exchange (TAIEX), and the experimental period is selected from 1997 to 2003 as the datasets for verifications. Finally, the experimental results showed that our proposed approach was effective in improving the forecasting errors on forecasting stock price significantly. Furthermore, the performances in terms of root mean squared error (RMSE) indicate that the proposed model is superior to the compared models suggested by Chen (1996), Yu (2005), and Chang et al. (2011) earlier. It is evident that the proposed model is a good approach to improve the forecasting performance fuzzy time-series models.