| Summary: | 碩士 === 國立臺灣科技大學 === 資訊工程系 === 94 === In this paper, we proposed twelve futures trading models to Taiex Stock Index Futures and explained the signals and results of these twelve models, and then we implemented two refiners, which were developed from two artificial methods¬¬--- Self-Organizing Map algorithm (SOM) and Fuzzy C-Means algorithm (FCM)---to revise the profit curves (PC) of these twelve models. We hope we can improve the profit curves and results of models to raise the efficiency of models.
After revised, we evaluated the strength and weakness of revised models by Profit Curve Refiner Index (PCRI), which includes Total Gain Ratio (TGR), Monthly Gain Ratio (MGR), Percent Profitable Ratio (PPR), and Psychological Ratio (PSYR). We tried to find the features of algorithm and the applicable model for the algorithm. Finally, refiners designed by SOM and FCM Algorithm can improve profit to common program trading model and raise Percent Profitable. It is indeed to increase the efficiency of models.
|
|---|
