MARKET SEGMENTATION WITH LATENT CLASS ANALYSIS : A STATISTICAL MODEL BY ANALYSIS OF CUSTOMER PURCHASE PATTERN

DOI Web Site 1 Citations 16 References
  • Sakurai Naoko
    Department of Environmental Information, Faculty of Informatics, Tokyo University of Information Sciences

Bibliographic Information

Other Title
  • 潜在クラス分析を用いたマーケットセグメンテーション : 顧客の購買パターンによるマーケットセグメント創出のモデル
  • センザイ クラス ブンセキ オ モチイタ マーケットセグメンテーション コキャク ノ コウバイ パターン ニ ヨル マーケットセグメント ソウシュツ ノ モデル

Search this article

Abstract

As the infrastructure of computer network has been greatly growing in the distribution business world, the data collecting function of frequent shoppers program typed data (abbreviated into "FSP") based on customer identification number has also been rapidly shown upward trend. When this kind of big sized database arranged time serial will be analyzed for customer segmentation, method of "target marketing" in which it will be segmented by customer purchase pattern data has been leading factor instead of current "mass marketing". In this paper it will be proposed a customer segmentation method of difference for homogeneous and heterogeneous by applying latent class analysis model based on purchasing pattern of each customer. It will be put the response frequency of separated brand as realized value from multinomial distribution and random variable for the frequency of purchasing brand products will be assumed as the Poisson distribution. Under such assumption, the parameter estimation for latent class has been done with EM algorithm. It is possible to characterize extracted each segment by the selecting probability of each branded product of belonging customer.

Journal

Citations (1)*help

See more

References(16)*help

See more

Keywords

Details 詳細情報について

Report a problem

Back to top