南京理工大学学报（自然科学版）
南京理工大學學報（自然科學版）
남경리공대학학보（자연과학판）
JOURNAL OF NANJING UNIVERSITY OF SCIENCE AND TECHNOLOGY
2014年
4期
531-536
,共6页
客户分类%购买时间%购买频次%平均购买额%购买倾向%K-means算法%初始聚类中心%聚类数
客戶分類%購買時間%購買頻次%平均購買額%購買傾嚮%K-means算法%初始聚類中心%聚類數
객호분류%구매시간%구매빈차%평균구매액%구매경향%K-means산법%초시취류중심%취류수
customer classification%recency%frequency%average monetary%trentd%K-means algorithm%initial clustering centers%number of clusters
为了解决传统K-means算法对初始聚类中心敏感和聚类数目事先难以确定的问题,提出了一种改进的K-means算法。改进算法利用最大距离等分策略来选取初始聚类中心,并利用一种评价函数来自动确定聚类数,减少了算法结果对参数的依赖。将改进算法应用到某企业客户分类中时,为提高分类结果的表征性,提出了以客户最近购买时间( Recency )、购买频次(Frequency)、平均购买额(Average Monetary)和购买倾向(Trend)作为客户价值细分变量的RFAT( Recency,frequency,average monetary and trend)模型,对客户RFAT值进行了聚类分析,并提供了针对不同客户群的营销策略。实证研究表明,该文所提出的改进算法和模型可以有效地对企业客户进行分类,能充分反映客户的当前价值和增值潜能。
為瞭解決傳統K-means算法對初始聚類中心敏感和聚類數目事先難以確定的問題,提齣瞭一種改進的K-means算法。改進算法利用最大距離等分策略來選取初始聚類中心,併利用一種評價函數來自動確定聚類數,減少瞭算法結果對參數的依賴。將改進算法應用到某企業客戶分類中時,為提高分類結果的錶徵性,提齣瞭以客戶最近購買時間( Recency )、購買頻次(Frequency)、平均購買額(Average Monetary)和購買傾嚮(Trend)作為客戶價值細分變量的RFAT( Recency,frequency,average monetary and trend)模型,對客戶RFAT值進行瞭聚類分析,併提供瞭針對不同客戶群的營銷策略。實證研究錶明,該文所提齣的改進算法和模型可以有效地對企業客戶進行分類,能充分反映客戶的噹前價值和增值潛能。
위료해결전통K-means산법대초시취류중심민감화취류수목사선난이학정적문제,제출료일충개진적K-means산법。개진산법이용최대거리등분책략래선취초시취류중심,병이용일충평개함수래자동학정취류수,감소료산법결과대삼수적의뢰。장개진산법응용도모기업객호분류중시,위제고분류결과적표정성,제출료이객호최근구매시간( Recency )、구매빈차(Frequency)、평균구매액(Average Monetary)화구매경향(Trend)작위객호개치세분변량적RFAT( Recency,frequency,average monetary and trend)모형,대객호RFAT치진행료취류분석,병제공료침대불동객호군적영소책략。실증연구표명,해문소제출적개진산법화모형가이유효지대기업객호진행분류,능충분반영객호적당전개치화증치잠능。
The traditional K-means algorithm has sensitivity to the initial cluster centers,meanwhile it is difficult for users to determine the optimal number of clusters in advance. In order to solve these problems,a new improved K-means algorithm is proposed here. The algorithm can optimize the initial center points through computing the maximum distance of objects. At the same time,it can find the optimal number of clusters by using a new evaluation function. The results can reduce the dependence on the parameters. When the improved algorithm is used to analyze customers of a firm, the RFAT customer classification model is proposed. The new model has four segmentation variables to assess the customer’s value:Recency, Frequency, Average Monetary and Trend. The customers RFAT-value is analyzed by using clustering. The business strategy for different customer groups is also pointed out. The application results show that the RFAT model and the improved K-means algorithm proposed here can classify customers effectively. It also can fully reflect the customer’s current value and appreciation potential.
