尋找關鍵品項購買序列以分辨 VIP 顧客之研究;Discovering Key Item Buying Sequences to Identify VIP Customers

NCUIR > School of Management at National Central University > Graduate Institute of Business Administration > Electronic Thesis & Dissertation > Item 987654321/70313

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/70313

Title:	尋找關鍵品項購買序列以分辨 VIP 顧客之研究;Discovering Key Item Buying Sequences to Identify VIP Customers
Authors:	林彥妘,;Lin,Yen-Yun
Contributors:	企業管理學系
Keywords:	資料探勘;RFM價值分析;K-Means分群;序列分析;Data Mining;RFM;K-Means;Sequential Analysis
Date:	2016-03-14
Issue Date:	2016-06-04 12:28:25 (UTC+8)
Publisher:	國立中央大學
Abstract:	根據80/20法則，企業有高達約80%的利潤僅來自於約20%的VIP顧客，如果能有效地分辨出這些少數的VIP顧客，企業就能夠集中資源服務重要顧客，減少重要顧客的流失率。其中，企業為了瞭解顧客需求以及消費行為，經常使用資料探勘(Data mining)的方法分析顧客的購買習慣，藉此針對不同顧客群制訂不同的行銷方案。　　本研究延續過去藉由關鍵品項尋找VIP顧客的研究，使用台灣某中型超級市場其中一家分店某年度之顧客交易資料，利用RFM價值分析法及K-Means分群將顧客分為VIP顧客及非VIP顧客，並使用序列分析且設定適當的最小規則支持度找出VIP顧客的頻繁品項購買序列，再以適當的門檻值篩選掉不具有鑑別力的序列，僅留下足以代表VIP顧客的關鍵品項購買序列。當一位潛在顧客購買了關鍵品項序列，則該顧客有很大的機會是為VIP顧客。　　本研究找到38組關鍵品項購買序列，驗證比率(Accuracy)約為78%。且因本研究僅需要可辨識的會員編號，故對於僅有會員編號的匿名資料探勘方法亦有相當的貢獻。 ;According to the 80/20 Principle, about 80% of the profits of a company comes from just 20% of customers. If we are able to identify these customers, companies can invest all the efforts to serve those customers and decrease their attrition rate. This important issue has been researched by only few studies. In these studies, data are treated as sets of buying baskets. In this study, transaction data are treated as sequence of buying baskets. The research then finding frequent item buying sequence through setting adequate minimal support in sequential analysis. Moreover, we choosing an adequate threshold value to filter only the key item buying sequences which have enough discriminability. This research discovered 38 key item buying sequence with the accuracy of 78%.
Appears in Collections:	[Graduate Institute of Business Administration] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	565	View/Open

社群 sharing

Loading...