本研究建立了整合品質屬性之重要度與滿意度調查的類神經網路之顧客群分類模式。經過品質屬性之重要度與滿意度調查及分析之後，再以因素分析來萃取出顧客所認為重要的品質屬性構面，並以這些重要構面來建立類神經網路學習模式。利用倒傳遞類神經網路具有辦識類別的特性，來建構顧客對於品質屬性之滿意程度的分群判別模式，並與多變量之線性區別分析和二次區別分析相比較。研究結果顯示，類神經網路的整體分類正確率和預測的精確性均較佳，顯示類神經網路有較佳的分類效果。此外，根據網路中輸入變數對輸出變數的貢獻度分析結果，可明顯得知影響顧客分群結果的決定性因素有那些。對公司而言，由分析的結果可以瞭解不同群體的顧客所需要改善的重點有那些，可作為擬定改善策略上的參考，對於提昇服務品質上有相當程度的貢獻。

This research establishes a classification model for customer groups using a neural network model. The proposed approach is based on the quality attributes from importance and satisfaction surveys. Factor analysis is utilized first to extract several important service quality attributes from customers. These attributes are used to construct the neural network learning process. The back-propagation neural network (BPN) is then used to establish a customer group satisfaction classification model. A similar classification model is produced using linear discriminant analysis (LDA) and quadratic discriminant analysis (QDA). The classification performances of these methods are then compared with that from the BPN model. After the comparisons, it can be verified that the overall classification rate and prediction accuracy of the BPN is superior to those obtained using the LDA and QDA models. According to the input node contribution analysis of the output nodes. we can obtain the determinant variables that affect the classified customer satisfaction groups. Therefore, companies can obtain improvement information from the contribution analysis to improve their service quality and increase customer satisfaction.