本研究納入消費者從事購買行為前的刪除和歸類行為現象，並藉由消費者所認知各個產品或品牌間的相似性，以模糊集群(Fuzzy Clustering)分析法中的模糊C 平均數(Fuzzy C-Means) 演算法建立競爭性市場結構(Competitive Market Structure) ，然後再根據消費者三選擇模式(Three choice E Model) 之資料，以最大概似估計(Maximum Likelihood estimation)法和非線性規劃求解參數估計，得到各品類或子市場忠誠者和潛在轉換者之轉換比率1藉由這兩種比率來分析競爭性市場結構，了解消費者行為，以作為研擬行銷策略之參考。實證研究中以臺南市轎車市場為研究對象，以年齡20至60歲有換購轎車經驗的消費者為研究母體，並得到三點結論：1. 基於消費者本身購車預算的影響，價格仍是消費者購車的第一取決要件。2. 消費者換購轎車時普遍有升級之考量。3. 各品類忠誠度不高，且潛在轉換者之重複替代購買比率為零。且從實證應用中，該亦發覺本模式所提供之資料將有助於正確地擬定行銷策略。

By employing the Fuzzy c-means of fuzzy clustering algorithms, this study implements the consumer's purchasing behaviors of elimination and categorization to construct a competitive market structure. The rates of switching and loyalty, based on the Three-choice Model of Consumers' behavior, are derived using the Maximum Likelihood Estimation and Nonlinear program. These rates can be used to analyze the competitive market structure to further understand the consumer's behavior, thereby facilitating the development of marketing strategies. The proposed model is used to analyze the purchasing behaviors of car owners in Tainan City (southern Taiwan). Those results are used to perform an empirical study in the proposed model's effectiveness is confirmed. Based on the results presented herein, we can conclude the following: Price is the critical factor of priority concern when potential customers determine which type they will purchase; Car owners tend to purchase cars that are more luxurious than their previous purchases; and Customer's loyalty to each type of car is extremely low. Furthermore, the likelihood that potential switching customers will purchase the same type cars which they currently own is zero. Results presented herein also demonstrate that the proposed model provides valuable information on how to enhance marketing strategies.

處於一個產品多樣化選擇的九〇時代，能成為贏家的公司，絕大都是那些善於分析其目標市場的需要，且精於確認機會所在，並能滿足目標顧客之最大需求者。不管是哪一種行業，其潛在的顧客群通常是遠大於既有的顧客群。因此，如何保有既有個顧客，亦即提高顧客忠誠度，以及發掘潛在顧客所在，增加市場佔有率，就成了個企業重視的一個課題。 為了分析公司所處的行銷環境，了解顧客的購買行為，行銷研究扮演了十分鐘要的角色。本研究採用模糊集群分析(Fuzzy Clustering analysis)和非線性規劃等數量工具來進行顧客忠誠度分析，依市場區隔變數中的人口統計變數，進行市場區隔分析，最後再分析個區隔中之購買者忠誠度狀態(Loyalty status)，及預估區隔間之轉換比率，以作為行銷決策之依據。 本研究適用於耐久產品市場中，對該種產品有購買經驗之消費者購買行為研究，經由分析各市場區隔顧客之忠誠度過程，以檢視購買型態背後之真正原因，瞭解攸關行銷面之SWOT，依此選擇正確的目標市場，有效的行銷策略也因此賴以建立。