Time Series Forecast Based on Grey Models

• Main Authors: Chai-Yen Wu, 吳佳諺
• Other Authors: none
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/34698587829575310599

碩士 === 國立臺灣科技大學 === 資訊工程系 === 96 === This paper predicts the time series based on the grey models. The value of the average line of moving was put into the grey models instead of the closing price. By taking the 5-days rolling MAP into the GM(1,1) and predicting the value of the next point from the time series, the result value is the predictive value of the average line of moving. The predictive closing value could be figured out by using the formula of MAF or RMAF. The 60-days average line of moving has the best predictive effects and the minimum error of the mean square. This proves MAP(60) is the best MAP in investment. The MAF was used for the trend forecasting, since the MAF represents the long term average line of moving. The crossing point of the closing price line and the long term MAF could be used to predict the variation of trend and to validate if the bear or bull market is coming. When the MAF(60) is above the closing price, it represents that it is the bear market now. When the MAF(60) is below the closing price, it represents that it is the bull market now. The RMAF can predict the more accurate closing price.