透過您的圖書館登入
IP:3.139.97.157
  • 期刊

市場購物籃資料關聯規則探測

On the Exploration of Localized Association in Market Basket Data

摘要


本研究使用關聯資訊來發現市場購物籃(market basket)項目間的關聯規則(association rule),或用來瞭解資料集(datasets)中,不同屬性間可能的局部關聯性,進而完成資料探勘的分類工作。局部關聯演算法能有效找出市場購物籃項目間特定的樣型(pattern)資訊,作為目標市場行銷之用。在所提出的方法中,利用客觀資訊測度規則,即可有效的同時找到多個局部正向與反向關聯聚類。另外,此方法也優於傳統市場購物籃分析法中,會發生處理時間隨資料集大小呈指數成長,及不易處理反向關聯規則的問題,因而在各種尋找局部關聯的應用上,將更具實用性。

並列摘要


In this paper, we propose a technique for discovering the item associations in market basket or understanding the localized associations in a multi-attributes categorical datasets using association information clustering method. Our work provide the algorithms which are more effective in discovering localized associations and can be used to perform the market basket analysis to find out the specific patterns which are useful for target marketing. Our method can simplify the time and space complexity problems and mining negative association rules simultaneously which are always encountered in the traditional support-confidence market basket analysis methods. Hence it's more suitable for real world localized association rules mining applications.

延伸閱讀