序列式因果關係之變化偵測模型

A Change Detection Model for the Sequential Cause-and-Effect Relationship

鄧任宏(Jen-Hung Teng)

指導教授 ： 許巍嚴 ； 黃正魁

國立中正大學/管理學院/資訊管理學系/碩士(2015年)

摘要

如何在動態環境中，識別行為模式的改變，一直以來都是個重要的議題，尤其因為資訊科技的快速進步，改善了舊有的資料收集方式和降低了資料的儲存成本，使得企業在利用資訊系統服務顧客的同時，也能輕鬆的取得大量的顧客行為資料，在現今，隨著資料探勘技術的發展，資料挖掘和分析的速度越來越快，企業越能從顧客行為資料中，發掘背後所隱藏的知識。
然而，當管理者從所收集的大量資料中挖掘感興趣的知識時，必須考慮一個重要的問題─在動態環境下，現存的知識是否一直適用?還是已經過時? 在本研究中，藉由序列分類樣式的挖掘和匹配，以達到幫助管理者判斷序列事件的因果關係之目的。如果管理者沒有即時地根據行為樣式的變化，進行知識的更新，那麼可能因為過時的知識而制定錯誤的決策，為了避免此一情況，管理者應該時常的更新他們的現存知識以對抗行為容易改變的動態環境。簡而言之，在動態環境中，為了回應持續發生的行為模式變化，管理者應該擁有理解行為變化知識的能力。
本研究利用Microsoft© SQL提供的樣本資料庫FoodMart作為研究資料集，在經過資料前處理後，將上述資料轉為序列分類樣式，並利用本研究提出的SeqClassChange架構，進行序列分類樣式的匹配和比較，在研究結果的部分，顯示了變化樣式的識別能輔助管理者進行決策上的幫助，因此我們相信本研究所使用的方法，能夠達到幫助管理者識別行為模式變化趨勢的目的。

關鍵字: 資料探勘、變化偵測、序列分類

關鍵字

序列分類 ； 變化偵測 ； 資料探勘

並列摘要

Identifying changes of customer behavior or event is an essential issue that must be faced for existing updating knowledge in a dynamic environment. Especially in nowadays, rapidly growth technology lets information collection becoming more and easier. Business can immediately collect numerous transactional data to discover the knowledge which is behind in their customers.

However, there is a problem− the knowledge which business uses data mining to be discovered with the data of customers is still suitable? In this study, we discuss a sequence-based classification pattern, which is used to figure out the sequential relation between cause and effect. The sequenced-based classification pattern may occur a situation that this pattern is suitable in the past time but is useless in nowadays. Without updating this knowledge, the manager will make an inappropriate decision. To settle this problem, this study proposes a novel change mining model, called SeqClassChange, to identify the change of patterns.

In the experiments, we use a FoodMart database which is stemming from the Microsoft© SQL Sample database. After the preprocessing procedure, we use our SeqClassChange model to get sequenced-based classification patterns, and then clarify the change of patterns. Experimental result shows how does change mining of pattern works; therefore, we believe that our method can help managers to identify the customer behavioral trends and to make a right decision.

Keywords: data mining, change mining, sequenced-based classification

並列關鍵字

Data mining ； Change mining ； Sequenced-based classification

參考文獻

Bala, P. K., Sural, S., & Banerjee, R. N. (2010). Association rule for purchase dependence in multi-item inventory. Production Planning & Control, 21(3), 274-285.

Chen, M.-C., Chiu, A.-L., & Chang, H.-H. (2005). Mining changes in customer behavior in retail marketing. Expert Systems with Applications, 28(4), 773-781.

Chen, Y.-L., & Huang, T.-K. (2005). Discovering fuzzy time-interval sequential patterns in sequence databases. Systems, Man, and Cybernetics, Part B: Cybernetics, IEEE Transactions on, 35(5), 959-972.

Chen, Y.-L., Kuo, M.-H., Wu, S.-Y., & Tang, K. (2009). Discovering recency, frequency, and monetary (RFM) sequential patterns from customers’ purchasing data. Electronic Commerce Research and Applications, 8(5), 241-251.

Cheung, K.-W., Kwok, J. T., Law, M. H., & Tsui, K.-C. (2003). Mining customer product ratings for personalized marketing. Decision Support Systems, 35(2), 231-243.