A neural network-based fuzzy time series model to improve forecasting

Neural networks have been popular due to their capabilities in handling nonlinear relationships. Hence, this study intends to apply neural networks to implement a new fuzzy time series model to improve forecasting. Differing from previous studies, this study includes the various degrees of membership in establishing fuzzy relationships, which assist in capturing the relationships more properly. These fuzzy relationships are then used to forecast the stock index in Taiwan. With more information, the forecasting is expected to improve, too. In addition, due to the greater amount of information covered, the proposed model can be used to forecast directly regardless of whether out-of-sample observations appear in the in-sample observations. This study performs out-of-sample forecasting and the results are compared with those of previous studies to demonstrate the performance of the proposed model.