基于马尔可夫链的间歇性购买个体客户终生价值预测模型

The Model of Predicting the Individual Always-a-share Customer′s Lifetime Value Based on Markov Chains Theory

随着市场竞争的日益激烈,客户的购买行为呈现出个体化和间歇性的特征,如何对此类间歇性购买个体客户的赢利性进行预测进而合理分配营销资源是当前亟待解决的一个重要问题。首先根据马尔可夫链的理论构造了间歇性购买个体客户的购买马尔可夫链,并利用马尔可夫链的转移矩阵得出该类客户全部剩余终生价值预测公式;其次,对终生价值预测公式中的参数(如个体客户在未来时间活跃度、个体客户在未来时间购买概率、个体客户在未来时间购买次数、个体客户在未来时间购买量等)进行了确定;最后,利用M atlab程序对济南市某IT分销商的客户交易数据进行了实证分析,从而证明了本预测模型的有效性和实用性。

With the increasing intensity of competition, the customer＇s purchase behavior gradually takes on the characteristic of individuality and intermission. In order to appropriately allocate marketing resource, companies should predict this kind of customer＇s profitability correctly. Therefore, it is important to know how to predict it. In this paper, firstly, based on the Markov Chains theory, the firm＇s relationship with the alwaysa-share individual customer is depicted. In addition, using Markov Transition matrix, the paper defines a model, which can predict all the residual profitabihty of the always-a-share individual customer. Secondly, the paper analyzes how to consider the parameter, such as alive rates, purchasing rates, purchasing times and purchasing amount. At last, using Matlab program and a IT company＇s dealing data, the paper carries on an empirical study for the purpose to testify the model＇s utility and validity.