基本分析是透過解讀財務資料來評估企業價值的一種方法,對一般投資者而言,基本分析中的財務報表是他們最容易取得,亦是最重要的參考依據,投資者可藉由企業每季發佈的財務訊息,去加以瞭解企業的經營績效和財務狀況,即使沒有實際參與公司整體運作,也可透過這些公開的資訊去評估投資該公司之報酬及風險,以預測企業應有的市場價值。 本研究採用演化策略結合類神經網路與Ohlson Model做為評估模組,但為了使整個模組更為完善,在Ohlson Model中非會計資訊部分加入公司治理相關變數,探討在何種Ohlson Model變數結構與數值範圍內,可以達到最佳的公司價值。 由於公司價值影響因素頗多,且因素之間呈現非線性關係,所以非一般傳統統計方法可以輕易解決,因此本論文採用人工智慧方法,藉由類神經網路良好的學習能力,加上演化策略擅於解決最佳化問題,在這兩大演算法的結合下,期望能找出三種類型公司的特性與範圍,並輕易確定公司股票的價值。 藉由本研究實驗找出的企業價值判別規則中之各變數的最佳範圍,可達到公司價值最大。如此能夠提供企業做為管理公司的依據,此外,亦可幫助投資者再投資股票理財時,可做為分析公司市場價值的一個參考。
Fundamental analysis is a method using financial data to evaluate the firm value. For common investors, the finical report of fundamental analysis is the most accessible and is the most important reference; the investors can through the financial information quarterly issued by the enterprise, know the operation performance and financial situation of the enterprise, so that even though the investors do not practically participate the enterprise’s overall operation, through such open information they can also evaluate the earnings and risk if investing in this enterprise, so as to predict the deserved market value of the enterprise. This research applies evolution strategy combined with neural network and Ohlson model to construct the evaluation module, but in order to make the whole module more complete, in the Ohlson model, corporate governance and etc. variables are added into the non-accounting information, in order to discuss in which variable structure and data range the Ohlson model can reach the optimal firm value. Because many factors influence the firm value and there is a nonlinear relationship between the factors, common traditional statistical method can not easily find solutions. Therefore, this essay applies the Artificial intelligence Method, which through the good learning ability of neural network and the good optimization ability of evaluation strategy, tries to work out the characteristics and ranges of three types of corporations and easily determine the corporate stock value. Use the firm value worked out by this research to identify the optimal rang of variables in the rules, thus to maximize the firm value. This can not only work as a reference for an enterprise to be a management company, but also can provide a reference for investors to analyze the corporate market value when again investing in the money management of stocks.