結合平衡計分卡與資料包絡分析法進行台灣發光二極體產業之績效評估

Abstract

本研究結合平衡計分卡(BSC)與資料包絡分析法(DEA)，建構分類模型以評估台灣發光二極體產業之績效。首先，以平衡計分卡為架構，利用四個構面進行初始變數選取，分別為財務、學習與成長、顧客及內部流程構面；接著利用統計方法進行變數篩選，並評估三種分類演算法之績效，希望能協助台灣發光二極體企業解決問題。本研究的目的有以下幾點：1.了解台灣發光二極體產業之關鍵變數，可視為績效差之公司優先改善變數2.比較變數篩選前與變數篩選後之分類模型的優劣3.探討營運效率與績效指標(股東權益報酬率與每股盈餘)間的影響。

This study combines the balanced scorecard (BSC) with the data envelopment analysis (DEA) and constructs classification model to assess performance for the Taiwanese light emitting diode industries. First, a balanced scorecard associated with four dimensions, such as financial, learning and growth, customer and internal process perspectives, is constructed. Second, using appropriate variable selection methods to search key variables. Finally, this study assesses the performance of three classification algorithms. The following is the purposes of this study: 1. The key performance indicators of the Taiwanese light emitting diode industries in order to improve poor companies. 2. Comparing performance of the classification models constructed by the original variables and the key variables. 3. Studying the relationship between the efficiency and performance variances (ROE and EPS).