基于Agent的超市客户关系管理系统的研究
发布时间：2019-06-19 13:26
【摘要】：随着经济发展和社会进步,超市已成为人们日常消费的主要渠道,如何在激烈的竞争中保持客户已成为各商家竞争的焦点。但目前超市行业在智能决策方面的研究尚处初级阶段,对存储在数据库中的数据也做不到深入的分析和挖掘。本文在智能决策理论的基础上对超市客户流失和客户价值进行了深入研究,建立了针对超市的基于Agent的客户关系管理系统。 本文通过客户流失预测模型和客户价值细分模型两个方面对如何制订即将流失客户的保持策略进行深入研究,论文完成的主要工作如下: (1)本文将Multi-Agent和数据挖掘技术引入超市客户关系研究,在JADE平台上实现了Agent成员间消息模式的通信机制,并通过各Agent间的相互协作,创建了基于Multi-Agent的超市客户关系管理系统。 (2)采用改进的C4.5决策树算法对客户进行流失预测,该算法在传统C4.5算法的基础上引入了加权因子并采用权重信息熵的计算方法,既可调节分类属性距根节点的位置又可使重要类别的特性更加突出,并通过理论及数据分析验证了该方法在超市客户流失预测方面的合理性和有效性。 (3)采用基于Agent的蚁群算法对预测为流失的客户进行客户价值细分,通过细分结果评定客户价值等级,以辅助超市管理者制定营销策略。Agent与蚁群算法的结合使该模块具有更强的鲁棒性,细分过程及结果也更加直观清晰,同时也提高了模型的自治性和社会性。 (4)整个系统按功能划分为客户流失预测Agent、数据中间处理Agent和客户价值细分Agent三个Agent成员,并通过JADE平台实现了各Agent间的通信及各模块间的智能化协调管理,提高了超市客户关系管理系统的自动化程度。 本文以超市客户关系管理为研究方向,详细阐述了系统的原理、实现过程及结果,从客户关系管理的角度对超市的客户保持策略进行了深入研究,为超市决策者制定保持客户的营销策略提供了决策依据,具有较高的理论价值和现实意义。
[Abstract]:With the development of economy and social progress, supermarkets have become the main channel of people's daily consumption, how to maintain customers in the fierce competition has become the focus of competition among businesses. However, the research on intelligent decision-making in supermarket industry is still in the primary stage, and the data stored in the database can not be deeply analyzed and excavated. In this paper, on the basis of intelligent decision theory, the customer loss and customer value of supermarket are deeply studied, and a customer relationship management system based on Agent for supermarket is established. In this paper, the customer loss prediction model and customer value segmentation model are used to study how to formulate the retention strategy of customers who are about to lose customers. The main work of this paper is as follows: (1) in this paper, Multi-Agent and data mining technology are introduced into supermarket customer relationship research, and the communication mechanism of message pattern among Agent members is realized on JADE platform, and through the cooperation among Agent, The supermarket customer relationship management system based on Multi-Agent is established. (2) the improved C4.5 decision tree algorithm is used to predict the customer loss. Based on the traditional C4.5 algorithm, the algorithm introduces the weighting factor and adopts the calculation method of weight information entropy, which can not only adjust the position of the classification attribute from the root node, but also make the characteristics of the important categories more prominent. The rationality and effectiveness of the method in the prediction of supermarket customer turnover are verified by theoretical and data analysis. (3) Ant colony algorithm based on Agent is used to subdivide the customer value of the customers predicted to be lost, and the customer value level is evaluated by the subdivision results in order to assist supermarket managers to formulate marketing strategy. The combination of agent and ant colony algorithm makes the module more robust, the subdivision process and results are more intuitive and clear, and the autonomy and sociality of the model are also improved. (4) according to the function, the whole system is divided into three Agent members: customer loss prediction Agent, data intermediate processing Agent and customer value subdivision Agent. The communication between Agent and intelligent coordination management among modules are realized through JADE platform, which improves the automation degree of supermarket customer relationship management system. Taking supermarket customer relationship management as the research direction, this paper expounds in detail the principle, realization process and results of the system, and makes a deep study on the customer maintenance strategy of supermarket from the point of view of customer relationship management, which provides the decision basis for supermarket decision makers to formulate the marketing strategy of maintaining customers, and has high theoretical value and practical significance.
【学位授予单位】：北京工商大学
【学位级别】：硕士
【学位授予年份】：2010
【分类号】：F274;F717.6
本文编号：2502375
[Abstract]:With the development of economy and social progress, supermarkets have become the main channel of people's daily consumption, how to maintain customers in the fierce competition has become the focus of competition among businesses. However, the research on intelligent decision-making in supermarket industry is still in the primary stage, and the data stored in the database can not be deeply analyzed and excavated. In this paper, on the basis of intelligent decision theory, the customer loss and customer value of supermarket are deeply studied, and a customer relationship management system based on Agent for supermarket is established. In this paper, the customer loss prediction model and customer value segmentation model are used to study how to formulate the retention strategy of customers who are about to lose customers. The main work of this paper is as follows: (1) in this paper, Multi-Agent and data mining technology are introduced into supermarket customer relationship research, and the communication mechanism of message pattern among Agent members is realized on JADE platform, and through the cooperation among Agent, The supermarket customer relationship management system based on Multi-Agent is established. (2) the improved C4.5 decision tree algorithm is used to predict the customer loss. Based on the traditional C4.5 algorithm, the algorithm introduces the weighting factor and adopts the calculation method of weight information entropy, which can not only adjust the position of the classification attribute from the root node, but also make the characteristics of the important categories more prominent. The rationality and effectiveness of the method in the prediction of supermarket customer turnover are verified by theoretical and data analysis. (3) Ant colony algorithm based on Agent is used to subdivide the customer value of the customers predicted to be lost, and the customer value level is evaluated by the subdivision results in order to assist supermarket managers to formulate marketing strategy. The combination of agent and ant colony algorithm makes the module more robust, the subdivision process and results are more intuitive and clear, and the autonomy and sociality of the model are also improved. (4) according to the function, the whole system is divided into three Agent members: customer loss prediction Agent, data intermediate processing Agent and customer value subdivision Agent. The communication between Agent and intelligent coordination management among modules are realized through JADE platform, which improves the automation degree of supermarket customer relationship management system. Taking supermarket customer relationship management as the research direction, this paper expounds in detail the principle, realization process and results of the system, and makes a deep study on the customer maintenance strategy of supermarket from the point of view of customer relationship management, which provides the decision basis for supermarket decision makers to formulate the marketing strategy of maintaining customers, and has high theoretical value and practical significance.
【学位授予单位】：北京工商大学
【学位级别】：硕士
【学位授予年份】：2010
【分类号】：F274;F717.6
【参考文献】
相关硕士学位论文 前10条
1 刘怡俊;电子商务系统中基于多Agent系统的信息流控制研究[D];广东工业大学;2002年
2 朱传宇;智能Agent及其在网络管理中的应用[D];合肥工业大学;2002年
3 杨玉凤;基于分析型客户关系管理的决策支持系统研究[D];广东工业大学;2004年
4 赵越;连锁超市企业营销战略与策略研究[D];首都经济贸易大学;2005年
5 王峥琦;基于决策树算法的改进与应用[D];西安科技大学;2005年
6 沈燕卿;基于数据挖掘的电信业客户流失分析与应用[D];重庆大学;2005年
7 李家新;会员制仓储超市客户关系管理的研究[D];吉林大学;2006年
8 王素霞;基于客户价值研究的客户关系管理[D];吉林大学;2006年
9 赵莽;基于实证分析的移动客户保持影响因素和策略研究[D];北京邮电大学;2006年
10 高海燕;基于数据挖掘的银行客户流失预测研究[D];西安理工大学;2007年
，本文编号：2502375
本文链接：https://www.wllwen.com/guanlilunwen/kehuguanxiguanli/2502375.html
