| Summary: | 碩士 === 國立臺灣科技大學 === 資訊工程系 === 94 === The purpose of this thesis is to design a trading system to reduce and lower investors’ psychological pressure.
In this research, a profit curve (PC) generated from a non AI-based trading model will be transformed into several technical indicators through three kinds of clustering algorithm: Grey clustering, Self-Organizing Map (SOM) and K-Means, therefore, the relative high and low points of PC are clustered. Moreover, investors may clear the original positions using signals of high, and reenter the market by the signals of low. Three profit curve refiners (PCRs): Grey Clustering Refiner (GCR), SOM Refiner (SOMR), K-Means Refiner (KMR) are further constructed. Finally, the characteristics of these three PCRs are discussed by using the evaluation metrics in this thesis.
|
|---|
