Summary: | 碩士 === 國立高雄應用科技大學 === 國際企業研究所 === 103 === According to various studies in existence, the discussions of influencing stock price were definitely separated into the basic component and the information entropy. Although a considerable number of portfolio analysis methods composed with basic components have been proposed, the discussion on methods with information entropy was still remained in blank. Interactive Artificial Bee Colony (IABC) in undeveloped artificial intelligent area was utilized in this thesis to propose and provide a proper set of quantified coefficients to evaluate the degree of influence caused by the occurrence of some important events in Taiwan broad stock market. The sliding window strategy is used in the experiment for assisting both training and testing phases. Based on the event study methodology, we focus on eleven economic and trade agreements signed between Taiwan and other countries in the history impacted Taiwan broad stock market being our study subjects. The selected agreements include Economic Cooperation Framework Agreement (ECFA), Asia-Pacific Economic Cooperation (APEC) and World Trade Organization (WTO), Taiwan-Japan Bilateral Investment Arrangement (BIA), Agreement between New Zealand and the Separate Customs Territory of Taiwan, Penghu, Kinmen, and Matsu on Economic Cooperation (ANZTEC), Agreement between Singapore and the Separate Customs Territory of Taiwan, Penghu, Kinmen and Matsu on Economic Partnership (ASTEP) and five FTAs with countries of Central America. In the experimental design, 70 days of pre-event period and 70 days of post-event period are selected for our observation period setting. The experimental results indicate that the forecasting error usually rise when the important event occurs. The influence is quantified and capable to be used for investors as a reliable tool to make better decision when investing in the Taiwan broad stock market. It might help and boost the domestic economic development as well.
|