A Genetic Rough Set Approach to Fuzzy Time-Series Prediction

Fuzzy Time series (FTS) has been widely applied to handle non-linear problems, such as enrollment estimation, weather prediction and stock index forecasting. FTS predicted values on the basis of an equal interval, which is determined the early stages of forecasting in the model. In this paper, we employed Genetic Algorithms (GA) to optimize the interval at first. Based on this, then Rough Set (RS) method is employed to recalculate the values. So the main purpose of this paper is to forecast a stock closing price by using the trend of the stock analyzed. We could identify trends of similar patterns from stock data to predict development of new data in the future. This proposed method is more efficient than the conventional FTS method.