摘要
提出了一个新的响应建模方法,这种方法更好地识别出那些对营销活动积极响应的客户,可以在数据挖掘的常用的有指导学习算法中实现,可通过逻辑回归的例子实现此方法。
This paper first puts forward the business objectives in database marketing and points out the current approaches toward the fulfillment of business objectives. It discusses the existent problems of the current approaches and then proposes a new methodology to identify the customers whose decisions will positively be influenced by campaigns. The proposed methodology can be used together with most commonly used and supervised learning algorithms. An example of Logistic Regression is used to illustrate the proposed methodology. At last, the procedure of estimation and validation models are presented.
出处
《桂林电子工业学院学报》
2005年第5期23-26,共4页
Journal of Guilin Institute of Electronic Technology
关键词
数据挖掘
数据库营销
预测模型
响应建模
模型评估
提升
data mining, database marketing, predictive model, response modeling, model evaluation, lift
