A Hybrid Genetic-Fuzzy Stock Selection Model Using Investor Sentiment Indicators

• Main Authors: Tsung-Nan Hsieh, 謝宗男
• Other Authors: Chien-Feng Huang
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/38234691176634059678

碩士 === 國立高雄大學 === 資訊工程學系碩士班 === 100 === In this thesis we present a study of hybrid genetic-fuzzy stock selection models using investor sentiment indicators. We first propose two basis strategies for the construction of stock selection models according to the degrees of optimism or pessimism of investors on the stocks. The genetic algorithms (GA) and fuzzy membership functions are then employed for optimization and flexibility of the models, respectively. In order to remove the constraint imposed by the pre-specified sentiment indicators, we further extend our models by using the GA to automatically determine the relationship between these indicators and future returns of stocks. Through our proposed stock selection models, the empirical results show that the model of buying pessimistic stocks outperforms the benchmark as well as the one of buying optimistic stocks. We also show that our proposed scheme for the GA-based free indicator model can further improve upon the two classes of the models using the pre-specified investor sentiment indicators.