Stock Turning Points Detection by Particle Swarm Optimization Clustering and Support Vector Machine Decision System

• Main Authors: Kai-yao Chi, 紀凱耀
• Other Authors: Chin-yuan Fan
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/797pgn

碩士 === 明道大學 === 產業創新與經營學系碩士班 === 98 === Stock price predictions suffer from some difficulties, i.e., non-stationary variations within the large historic data and uncertain financial and political situation. This paper establishes a novel stock investment decision system by Particle Swarm Optimization clustering and Support Vector Machine decision for stock price movement predictions and investment decision in Taiwan famous industries stocks. This forecasting model integrates Particle Swarm optimization technique to clustering stocks data into several groups, and Support vector Machine to construct a decision-making system based on historical data and technical indexes. The model is major based on the idea that the historic price data base can be transformed into a smaller cluster together with a group then the Support vector Machine model can be more accurately react to the current tendency of the stock price movement from these smaller cluster group inductions. Hit rate is applied as a performance measure and the effectiveness of our proposed PSO-SVM model is demonstrated by experimentally compared with other approaches in various stocks from Taiwan Stock Exchange Center (TSEC).