马尔可夫链概率矩阵的银行业客户行为预测

摘要

对客户交易行为的预测是企业营销的重点问题.一种基于马尔可夫链状态转移矩阵的客户行为预测方法,利用NBD模型计算客户在不同状态下的下一次购买概率,并根据银行业的特点对模型进行了改进,利用客户新鲜度和交易级别的变化反映客户特征.最后利用数据对模型进行验证,详细说明了计算方法及其现实的意义.%The forecasting of customer' s next purchase is a focus of concern in marketing. A detailed introduction is made of a method for forecasting customer's next purchase based on state-transition Matrix of Markov Chian. The NBD model was used to calculate the probaility of customer' s next purchase under different conditions. Then the model was improved according to the characteristics of banks. Finally,the model was verified and its calculation and significance were discussed in detail.