在資訊科技的推波助瀾下,不僅企業競爭的強度與速度倍數於以往,激增的市場交易也使得各企業所需儲存與處理的資料量越來越龐大。為了因應外界的競爭,如何能快速且有效的從資料庫中取得有用的資訊,並反應市場或消費者的需求,成為各企業重視的焦點。藉由資料探勘的技術可從龐大的交易資料中找出顧客的消費特徵,並透過完善的顧客關係管理策略的運用,來提高顧客的忠誠度與企業之獲利。 本研究將國內某金融機構之顧客資料庫的顧客基本資料及刷卡記錄,利用類神經網路中的自組織映射圖網路 (SOM) 與統計集群分析的K群平均數法 (K-means) 兩種分群方法做比較,找出適用於發卡銀行之顧客資料庫的區隔模式,並根據資料探勘過程中所獲得的相關資料進行統計分析,最後應用各種行銷手法於不同的顧客類型,以提升顧客之忠誠度,增加發卡銀行的競爭力。
Because of the information technology expanding and the keen competition among the financial institutions, the miscellaneous consumers’ data lead to data processing hardly in enterprises. Therefore, how to obtain the useful information from the consumers’ database to find out the their demand becomes the most important target of the financial institutions. In addition, using the hot technology, data mining, can find out the characteristics of consumers’ behavior from the diverse database and establish the fit strategy for the consumer relationship management. This research will compare self-organizing map (SOM) with the statistical cluster analysis, K-means, by analyzing the consumers’ database of the financial institution. We obtain the discriminative model for the consumers’ database of the card-issuing bank and utilize the different marketing modes for each distinct consumer’ group to promote the consumers’ loyalty and strengthen the competitive ability of the card-issuing bank.