基于GMDH动态聚类集成的应用商店客户价值细分研究

项目名称： 基于GMDH动态聚类集成的应用商店客户价值细分研究

项目编号： No.71501136

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 管理科学

项目作者： 腾格尔

作者单位： 四川大学

项目金额： 17.4万元

中文摘要： 手机应用商店已成为移动互联网未来的发展趋势。无法辨别客户群体的需求是应用商店当前发展面临的瓶颈，客户价值细分则是解决该问题的有效途径。影响应用商店客户价值细分的诸多因素中，应用程序免费和客户样本数据持续增加是最主要因素。考虑应用商店的独特性，本项目对应用商店客户价值细分的理论与方法进行深入研究。具体内容包括：提出了应用商店客户价值细分研究框架。在该框架下，针对应用商店应用程序免费的特点，通过客户价值比照实验确定客户价值细分指标；针对应用商店客户样本数据的持续性特点，以数据分组处理方法（Group Method of Data Handling，GMDH）自动建模机制为基础，提出两种动态基聚类器生成策略，并构建两种基于GMDH的动态聚类集成模型用于客户细分。研究成果对推动应用商店发展具有重要意义，为客户价值细分和动态聚类集成奠定理论和技术基础。

中文关键词： 应用商店；客户价值细分；GMDH；动态聚类集成

英文摘要： Mobile application stores have become the trend of the future of mobile Internet. Unable to distinguish the customers’ demand has become the bottleneck of mobile application store. Customer value segmentation is an effective way to solve such problem. Two factors are the most influential factors that affect the customer value segmentation in the mobile application store: most of applications are free and the data collected from customers continue to grow. By considering characteristics of the mobile application store, this proposal researches the customer value segmentation theories and methods used for the mobile application store. A customer value segmentation research framework is proposed. Then, an experiment is designed to determine customer value segmentation index by considering the characteristics of the free applications. By considering the features of the collected customer data, this proposal investigates two kinds of dynamic base clustering generation strategies based on the Group Method of Data Handling (GMDH). Finally, two dynamic clustering ensemble models based on GMDH, which used for customer segmentation, are also studied. This proposal has great significance to promote the development of the mobile application store. It also lays a foundation for the theory and algorithm of customer value segmentation and dynamic clustering ensemble.

英文关键词： Mobile application store; Customer value segmentation;GMDH;Dynamic clustering ensemble