投資者可以使用公開的財務季報表去了解企業的營運績效,根據這些資料,本論文應用演化策略結合類神經網路和Ohlson模型去建構一評估模組。因為有許多因素會影響公司價值,且這些因素之間存在著非線性和複雜的關係,所以本研究利用人工智慧方法,透過類神經網路的良好學習能力,與擁有最佳化優點的演化策略,試圖找出三種公司類型的特性和變數範圍,並可以輕易評估其相對的股票價值。本研究所找出企業價值評估規則的變數最佳範圍,其正確率高達90%以上,並可以最大化公司的價值。本論文所提出的方法不僅可以提供給企業管理參考,更可以幫助投資者分析公司的真實市場價值。
Investors can use quarterly financial reports issued by enterprises to know enterprise operating performance. This study applies evolution strategies combined with the artificial neural network and Ohlson model to construct the evaluation module. Because many factors influence firm value and the relationship among those factors are nonlinear and complex. Therefore, this study applies the artificial intelligence method, which through the good learning ability of artificial neural networks and the good optimization ability of the evolution strategies, tries to identify the characteristics and their ranges of three types of corporations and easily value the related stock. This study uses the calculated firm value to identify the optimal range of variables in the rules, their accuracy get over 90%, and thus to maximize firm value. This approach can not only provide a reference for enterprise management, but also one for investors to use in analyzing corporate market value.