財務報表分析一直以來是大家關注的議題,因為財務報表隱含了相當多影響公司營運的指標,也衡量企業在某一期間的財務狀況和經營成果。然而分析財務報表是耗時耗力的非結構性工作,應藉由系統輔助,獲得分析結果,減少人為的錯誤決策。本論文之目的在利用基因演算法學習專家行為,建立一套基於XBRL之財務報表分析。XBRL易於網路使用的特性,可以使提供者輕易地將其公司的財務報表、分析結果等提供給使用者。本研究以分析長期償債能力為例,利用基因演算法來調整財務指標的權重,並驗證是否可以達到和專家相似的分析結果,讓一般非財務背景的使用者也能受惠,加上圖形化介面,使用者亦可以根據不同需求動態的更改知識及等級區間的範圍,以獲得更佳的分析結果。
The analysis of financial statement has been a public concerned issue. It not only includes many indices which influence the management of enterprise but also is a result of company operation. However, the analysis of financial statement is an unstructured work which costs time and money. Therefore, to establish an auxiliary system would be useful to assist the decision-making and minimize human errors. The purpose of this research is to establish a financial analysis system based on XBRL (eXtensible Business Reporting Language). In this study, the weight of different indices would be adjusted through genetic algorithm to analyze the debt-paying ability in long terms and verify whether the analysis results could fit the professional results. The developed system benefits users without financial background. With the GUI users also can dynamically alter the knowledge and evaluation level according to different needs to obtain better analysis result.