隨著商業經濟全球化及資訊科技快速發展下,企業透過資料倉儲建立龐大的顧客及交易資料庫,但如何從龐大資料挖掘出潛在的資訊與規則,已成為現今大數據時代企業競爭的成功關鍵因素。而龐大資料中所隱藏的資訊與規則可透過資料探勘技術來挖掘萃取,結果可運用於建立顧客關係、協助企業營運、提升決策品質等用途。 顧客(直銷商)是直銷公司最核心的經營要素,因此如何經營與顧客的關係,以達到企業永續經營的目標,是直銷公司關鍵的經營課題。因此本研究選擇直銷公司為個案進行探討,以個案公司的顧客交易資料庫資料,首先使用 RFM分析法,產生R、F、M三項指標衡量值作為輸入值,再使用K-means集群分析法對顧客進行分群,依顧客分群結果提出顧客關係經營策略建議。最後依據R、F、M三項指標衡量值與顧客分群結果,以C5.0決策樹演算法產生模型與規則組集,瞭解顧客群體之特徵,提供後續顧客價值的預測參考,協助企業深化顧客洞察能力,進而增進企業收益。
According to business economic globalization and the rapid development of information technology, enterprises establish a huge database of customers and transactions through the data warehouse. How to dig out the potential information and rule from big data has becomes the critical factor that drives enterprises competition to be success in the age of big data. Potential information and rule in big data can be dug out through data mining technology. It can be used in establishing customer relations, assisting business operations, improve the quality of decision-making and other purposes. Customers (multilevel marketing sales) are the most heart of multilevel marketing operating elements. How to manage with customer’s vertical relation. In order to achieve the goal of enterprise tsustainable operation is multilevel marketing company’s key operation issue. The research chooses multilevel marketing as a study case to discuss.The example of company’s customer data of transaction database. In the first, use RFM analysis to get R, F and M three index rate as input value. Next, use K-means clustering to classify customers. According to the results, presented customers relationship management strategy recommendations. At last, according to R, F and M three index rate and the result of customer classification, produce models and set of rules with C5.0 decision tree algorithm. Understanding of the characteristics of customer groups, provides customer value prediction reference, assisting enterprises have an insight into customers deeply, and enhance enterprise earnings.