本研究核心目標是發展新式的多準則決策模型。此模型預設決策者在複雜多元的資訊環境中，先藉由約略集合為核心的演算法，淬取出具代表性的多重法則；再讓決策者依據本身重視非典型觀測值的程度，決定是否逐步挖掘較深層（由特殊、相對少數觀察值所組成）的法則，稱為「漸進探索式多法則決策模型」。其次，將此模型應用在評估上市櫃企業的公司治理績效，歸納出評估公司治理績效的多重法則。延伸目標是藉由此模型所得到的公司治理法則，轉為應用在證券化代幣募資（Security Token Offerings, STO）的數位金融市場，協助交易所評估STO專案的公司治理績效，進而提升台灣發展STO市場的國際競爭力。
The essential goal of this research is to develop a new type of multiple rule-based decision-making (MRDM) model. This model assumes that decision makers are in a complex and diverse information environment. The model leverages machine learning algorithms (based on rough sets) to extract a representative group of rules, supported by typical observations. In the next, the decision makers may decide whether to explore deeper (composed of relatively few observations) logical rules based on how much they value non-typical observations. Secondly, this decision model will be used to evaluate the corporate governance performance of listed companies, and the multiple rules that can evaluate corporate governance will be induced. The extended goal is to transform the obtained rules for evaluating the corporate governance of Security Token Offerings (STO) projects. This will further enhance Taiwan's international competitiveness in developing STO digital financial markets.
