台灣光電產業經營績效評估-結合DEA與DB-SCAN群聚分析

Abstract

資料包絡分析模式(DEA)被廣泛應用於各行各業之「經營績效」分析，對受評公司作排名。群聚分析法(Clustering)廣泛應用於找出相似性質的群組，並進ㄧ步分析群組之共同特性。 本論文根據企業經營的評比條件，以DEA對光電產業公司的「經營績效」作排名(ranking)，並以DB-SCAN群聚分析法，區分公司經營績效等級，並進ㄧ步分析各個群組等級的特性。本論文所提出之產業競爭力分析模式，可協助台灣光電產業公司，了解企業在產業群聚的相對位置與優勢。

關鍵字: 包絡分析法(DEA), DBSCAN, 群聚法,光電產業, 績效評估

Data envelopment analysis (DEA) is a widely used technique to rank the performance of businesses in performance evaluations. Clustering is widely utilized to find clusters with similar features and analyze their common characteristics. This thesis uses DEA to rank the performance of optronics companies based on the evaluation criteria of business performance. Then, DB-SCAN clustering is applied to cluster optronics companies into several performance groups. The common features of each performance group is derived and analyzed. The proposed competence analysis approach can help businesses understand their strengths and positions in the industry. Keywords: DEA, DBSCAN, Clustering, Optronics Industry, performance evaluation