文章詳目資料

管理資訊計算

  • 加入收藏
  • 下載文章
篇名 台灣大型集團企業經營績效之財務指標採礦研究
卷期 3特刊
並列篇名 Data Mining of Financial Indicators for Large Enterprise Groups in Taiwan
作者 林雅俐劉淙凱
頁次 480-489
關鍵字 資料採礦財務指標大型集團企業預測模型資產報酬率Data MiningFinancial IndicatorsLarge Enterprise GroupsPredicted ModelReturn on AssetsROA
出刊日期 201410
DOI 10.6285/MIC.3(1)S.39

中文摘要

本研究針對台灣「集團企業排名」資料庫(中華徵信所,2012)299個大型集團企業(區分為24個產業)之十項財務指標,探討影響企業經營績效之關鍵因素。使用2006~2012年共1919筆資料,選取資產報酬率(ROA)作為企業經營績效之評估標的(Hitt等人,1997)。研究結果顯示,決策樹模型和邏輯斯迴歸模型可以有效建構資產報酬率之預測模型,其中淨值報酬率和自有資金比率為預測資產報酬率盈虧機率之最佳財務指標,將此模型命名為成長率資產模型,其中決策樹成長率資產模型於訓練、驗證、和測試資料下所得之正確預測分類比率分別為98.3%、97.7%、和89%,同時邏輯斯迴歸成長率資產模型於訓練、驗證、和測試資料下所得之正確預測分類比率亦達91.1%、94.6%、和86%,顯示成長率資產模型可作為預測資產報酬率盈虧機率之最佳預測模型。

英文摘要

Based on Enterprise Group Data base of China Credit Information Service, Inc. in 2012, 299 Taiwanese large enterprise groups and 24 industry sectors were extracted in this study. Return on assets (ROA) is used to evaluate the operational performance as the target variable which is as same as the research of Hitt, et al. (1997). The results indicate return on equity (ROE) and self-owned capital ratio are statistically significant in the decision tree and logistic regression models. The rates of correct classification reach as 98.3%, 97.7%, and 89%, respectively, based on the training, verification, and testing data for the decision tree model. In addition, the rates of correct classification are 91.1%, 94.6%, and 86%, respectively, based on the training, verification, and testing data for the logistic regression model. The growing rate and assets model is the best one to predict the likelihood of profit or loss of ROA in this study.

本卷期文章目次

相關文獻