首页 | 本学科首页   官方微博 | 高级检索  
     检索      

一种智能学习的动态客户标签方法的研究
引用本文:苏燕,;王猛,;苗润生.一种智能学习的动态客户标签方法的研究[J].铁路计算机应用,2014,23(8):16-19.
作者姓名:苏燕  ;王猛  ;苗润生
作者单位:哈尔滨铁路局信息技术所,哈尔滨,150008%哈尔滨铁路局客运处,哈尔滨,150008
摘    要:研究客户标签的生成机制,构建一种动态的客户标签更新方法,结合One-Class分类和支持向量机算法智能模拟人类的学习行为,能够在模型的复杂性和学习能力之间衡量并不断自我优化,仅通过典型有限的小样本训练集,即可建立客户标签模型,又由于小样本可有效提高运行效率,实现客户标签的动态更新.

关 键 词:客户标签    向量机    One-Class    动态
收稿时间:2015-10-09

Dynamic customer label method of intelligent learning
Institution:SU Yan,WANG Meng,MIAO Runsheng( 1.Institute of Information Technology, Haerbin Railway Administration, Haerbin 150008, China; 2.Department of Passenger Transport, Haerbin Railway Administration, Haerbin 150008, China)
Abstract:This article mainly studied on the mechanism of the customer label,built a dynamic customer label update method,combined with One-Class classification and support vector machine (SVM) algorithm to simulate human learning behavior intelligently,was able to measure between the model's complexity and learning ability and be constant self optimization.The customer label model could be built by typical limited small sample training set.The small sample could effectively improve the operation efficiency,implement a dynamic update of customer labels.
Keywords:customer label  vector machine  One-Class  dynamics
本文献已被 维普 等数据库收录!
点击此处可从《铁路计算机应用》浏览原始摘要信息
点击此处可从《铁路计算机应用》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号