運用類神經模糊模型探索策略配適與企業績效之關係

Using Neuro-fuzzy Model to Explore the Relationship between Strategic Fit and Firm Performance

本研究嘗試以策略配適的觀點來討論企業績效的優劣，利用產業組織學派以及資源基礎觀點，將企業的配適內容劃分為外部與內部配適，利用輪廓離差模型來探討外部與內部配適對企業績效的影響。以台灣上市櫃電子企業在2003年至2009年間的追蹤資料為研究對象，透過追蹤資料迴歸模型加以分析。實證結果顯示，外部與內部配適程度越高時，企業績效確實會越好，然而，外部與內部配適對企業績效的影響並未呈現顯著的交互效果。為了進一步挖掘外部與內部配適對績效可能存在的交互作用影響，本研究進一步利用類神經模糊工具，以2003年至2007年間的樣本來建構變數間可能存在關係的知識庫，並以2008年至2009年的資料來測試所得知識庫的正確程度。實證結果顯示，透過類神經模糊所得到的知識庫，除了可以提供業界在策略配適決策時重要的決策意涵之外，也呈現出外部與內部配適對企業績效的影響，存在著迴歸交乘項之外的非線性關係。對於社會科學中，變數之間可能存在的非線性複雜關係，類神經模糊確實提供了一個可能的解決之道。

This research is trying to explore the firm performance from the strategic fit perspective. Based on the contents of Industrial Organization School and Resource Based View, this research divides the strategic fit of a company into external and internal fits, and explores how these fits affect the performance by utilizing the profile deviation approach. With panel regression technique applied to the electronic enterprises listed in the Taiwan's Stock Market from 2003 to 2009, the empirical results derived show that the higher the external and internal fits, the better the performance. However, the empirical results do not show a significant effect on performance from the interaction term between external and internal fits. This research applies neuro fuzzy further to explore the possible effect on performance from the interaction term between external and internal fits. The data set within the interval from 2003 to 2007 is used to construct the knowledge base among these variables, and the data from 2008 to 2009 is used to verify the model validity derived. The empirical results show that the knowledge base not only can provide the decision makers the practical implications, this technique also show the possible effect on performance from the interaction term which is undetected by the panel regression model. This technique can be a possible solution to problems where there exists a complicated relationship among the variablesg.