| 摘要: | 在過去近十年中，學者發表許多有關企業社會責任與財務績效的研究，多以較傳統的迴歸模型去分析，資料探勘技術鮮少用在此研究主題，因此本研究以資料探勘方法，探討企業社會責任與財務績效之關聯性。本研究採用中國大陸企業社會責任藍皮書，根據書中針對企業社會責任的「責任管理」、「市場責任」、「社會責任」、「環境責任」作為衡量企業社會責任績效的指標，有別於以往除了探討整體績效指標外還有探討四個指標對財務績效之影響。
本研究以每股盈餘、資產報酬率與股東權益報酬率作為財務績效代理變數，分別採用約略集合理論(rough set theory)及決策樹(decision tree)加以判別，實證結果顯示以每股盈餘作為財務績效代理變數之準確率最高，其準確率大部分皆達80%以上；至於影響財務績效的重要變數，在企業社會責任變數中以責任管理相對較為重要，非企業社會責任變數則以企業成長率與企業規模相對較為重要。
In the past few decades, scholars have published many studies on corporate social responsibility and financial performance, more to the more traditional regression model to analyze, data Mining technology rarely used in this research topic, so this study using data mining to explore the relationship between Corporate Social Responsibility and Financial Performance. According to the Research Report on Corporate Social Respon-sibility of China, this study is used "Responsibility Management", "Market Responsibil-ity", "Environment Responsibility" as a measure of corporate social responsibility per-formance. In addition to exploring the overall performance indicators, there are four in-dicators on the impact of financial performance.
In this study, earnings per share, return on assets and return on equity as financial performance proxy variables. Respectively, using rough set theory and decision tree to be identified. The empirical results show that the earnings per share is the highest as the financial performance agent variable, the accuracy rate of most of them are more than 80%. As for the important variables that affect financial performance, it is relatively important to manage responsibility in corporate social responsibility variables. Growth rate and the Firm size is relatively important in non-corporate social responsibility var-iables. |
