摘要
首先简述了一个线性市场值函数模型,然后对其中的关键技术进行了讨论,设计并实现了一个挖掘市场值函数的原型系统,最后给出了该系统在具体数据集上的试验结果。试验结果表明,在该系统上使用Lift评价方法计算出来的Lift值在66%左右,具有一定的实用性;与其他使用分类方法的系统相比,该系统不仅能够根据市场值函数的值来排列对象,而且还能向使用者提供属性的重要性等信息,从而有利于更好地对顾客群体进行识别,对目标市场进行分析。
The authors first briefly illustrate a linear market value function model, then discuss its key technique, design and implemnt a prototype system of mining market value function, and finally give the experimental results on the system's special data set. The results show that the Lift value calculated by the system with Lift evaluation method is about 66% and has a cartain practicability. As compared with other systems adopting other classification methods, this system can not only arrange customers according to the market values but also provide the more important attributes so as to identify customers better and analyze the target market.
出处
《北京工业大学学报》
CAS
CSCD
北大核心
2003年第1期94-97,共4页
Journal of Beijing University of Technology
基金
国家自然科学基金资助项目(601730014)
北京市自然科学基金资助项目(4022003)
北京市教委科技发展基金资助项目