本研究旨在探討台灣銀行產業之經營績效分析,研究期間為2005年至2008年,以國內21家廠商為樣本,透過資料包絡分析法 (data envelopment analysis,DEA)及整合DEA與類神經網路模型(artificial neural networks,ANNs)。其先在不考慮其權重及環境變數等因素,採用一般DEA模式進行經營績效分析。接著分別採用資料包絡法與整合資料包絡法與類神經網路模式進評估台灣銀行產業經營效率之比較,亦藉此探討影響經營效率之技術效率、純技術效率及規模效率之差異性。 根據實證結果發現一般DEA模式下之平均效率值均為較高,而一般DEA模式透過整合類神經網路方法下之平均效率值均為較低,確實可以增加效率結果之差異性,以表現出台灣銀行產業之真實經營效率。基本上,這些分析所獲之資訊,大致能對台灣銀行廠商在未來經營效率提升之策略擬定與政府政策擬定之參考。
This research aims to examine Taiwan banking industry performance analysis, study for 2005 Years to 2008 Years in Taiwan 21 Home manufacturers for samples through the data envelopment analysis method (data envelopment analysis,DEA) And integration DEA And neural network models ( artificial neural networks,ANNs )。 Precedent without considering their weights and factors such as environment variables,using DEA model for performance analysis. Then using datagram collaterals and integration of data by envelope method and neural network model into evaluating Taiwan banking industry efficiency comparison, would like to take this discussion affect technical efficiency, pure technical efficiency of the operational efficiency and scale of efficiency difference. Based on the empirical results found using DEA model Average efficiency of the mode value is high, and the DEA Mode through the integration of neural network method of average efficiency values are low, does increase efficiency result of difference, to show real efficiency of the Taiwan banking industry. Basically, the analysis of the information, generally on Taiwan banking vendors in future operating efficiency enhancement of policy formulation and the reference of Government policy formulation.