在日益競爭的市場中，若要比其它競爭者獲得更多之優勢與先機，對於消費者行為不可忽視。因此，在銷售資料庫中如何有效地挖掘出有用的規則，將有助於企業決策時有價值的資訊。本研究結合螞蟻路徑尋優法與主成份因素分析，發展出一套整合性資料挖掘方法，將可有效找尋出可能的商品關聯規則。研究方法中同時考慮賣場商品佈置與銷售資料所求得為有效率路徑式商品關聯，這將可有效克服過去研究中，大致上以兩者商品為主的關聯規則，因為兩者以上商品將會產生規則數眾多而失去意義，以及因為調整閥值而導致關聯規則變動的缺點。研究結果顯示此方法能有效地找尋與分析隱藏在資料庫中的知識，進而提供給決策者更有用的資訊，例如配合賣場佈置與銷售資料進行瞭解消費者購買商品的行為，使得未來在擬定賣場商品陳設、商品選配或是促銷方案，有更好的參考資訊。研究中亦利用從一大型量販店所收集的賣場佈置與銷售資料為範例，並配合其他資料挖掘套裝軟體XpertRule Miner與本研究所建立的整合性資料挖掘模式的結果作分析比較。

Customer behavior is not to be ignored in order to gain more advantages than other competitors in an increasingly competitive market. Therefore, effectively extracting useful patterns in database would be beneficial for companies to discover practical knowledge in decision making process. This research is to develop a combination data mining method by integrating the concept of Association Rules, along with path finding and Principal Component Analysis in Ant Theory. By taking both of floor layout and sales data, this research is expected to find flow type product association rules, in turn, it can reduce the drawbacks of past research such as the mainly two-product association rules and the too many number of possible rules. Research results show that the developed method can effectively locate and analyze knowledge hidden in the database, and to provide more useful information to decision makers. For example, recognizing consumer purchasing behavior provides more precise references in future designs of promotion campaigns, product selections and product displays in the store. This study also compares sales data obtained from a large chain store processed in data mining software, XpertRule Miner, to results from the integrated data mining model.