隨著資訊科技的蓬勃發展下，使得企業間的競爭態勢比以往更為激烈，因此，企業必須能夠確認不同顧客群的目標市場，針對不同的顧客群給予不一樣的行銷策略，如此方可提高企業的競爭優勢。在現今市場中，企業以商品組合方式銷售之行銷策略，已普遍運用。若成功將這些商品組合推薦給合適的目標顧客群，將可達到吸引顧客及提高銷售量之目的。在這樣的環境下，資料探勘已成為企業從資料庫中有效挖掘出消費者的習性以及需求的技術。因此本研究目的在於利用資料探勘技術提供管理者針對顧客的交易行為進行分析，作為顧客關係管理的參考，以利於行銷策略之運用。 本研究之方法是利用Apriori演算法挖掘出商品間的關聯性，根據其關聯性給予組合方式銷售。然後透過改良後之模糊約略集合理論（Fuzzy Rough Sets Theory）找出顧客特徵與商品組合間的關聯性，進而設計出一套系統，最後透過一些實驗來驗證改良式之模糊約略集合所產生法則之精確性及有效性。 In the drastic competition environment, it is an important task that the enterprise must give different marketing strategies based on the different targeting customers to raise its competitive advantages. In the present market, the marketing strategy of the sale by merchandise combination is in widespread use by the enterprise. The enterprise can gain a lot of profit from recommending the relevant merchandise combinations to the targeting customer to promote the consumption desire of the customers. With the help of the knowledge discovered from the databases, the enterprise can effectively understand the consumer's buying behaviors and requirements. In this study, we propose a model to carry on the market segmentation, according to the characteristics of the customers and their consumption behaviors, so that the business managers can find out the better marketing model. In our proposed model, the Apriori algorithm is first used to mine the frequent merchandise combinations from the customer transaction database. Then the relationships of the customer characteristic and merchandise combinations are discovered through the Fuzzy Rough Sets Theory. Several modifications are proposed to mine the more useful rules. Some experiments are conducted to evaluate the feasibilities and effects of the proposed system.
