期刊文献+

基于深度置信神经网络的电信客户流失分析

下载PDF
导出
摘要 电信客户流失预测难度较大,因为针对的是人的行为问题,人的行为琢磨不定,因此需要多种基于观察人的数据特征,包括行为特征和基本特征,通过对这些特征建模来解决问题。针对电信行业客户流失的问题,本文设计提出一种基于(XGB)极度梯度提升树和深度置信网络(DBN)的电信客户流失预测方法(XGB-DBN)。首先利用XGB对原数据集产生新特征,通过和原数据集的集成,再通过深度置信网络对新数据集进行学习建模,对某电信公司宽带客户流失数据进行了仿真。实验表明,XGB-DBN获得的召回率、精确率、准确率、F1分数远远高于其他预测方法,说明极度梯度提升树结合深度置信网络的数据挖掘方法具有很好的预测效果,为电信客户流失预测提供了一种新方法。
出处 《通讯世界》 2020年第6期189-190,共2页 Telecom World
  • 相关文献

参考文献2

二级参考文献15

  • 1PU C, WEBB S. Observed trends in spam construction techniques:a case study of spam evolution [ C]// CEAS 2006: Proceedings of the Third Conference on Email and Anti-Spare. Mountain View, California: CEAS, 2006:104 - 112.
  • 2PUNIKIS D, LAURUTIS R, DIRMEIKIS R. An artificial neural nets for spam E-mail recognition [ J]. Electronics and Electrical En- gineering, 2006, 69(5): 73-76.
  • 3ANDROUTSOPOULOS I, PALIOURAS G, MICHELAKIS E. Learn- ing to filter unsolicited commercial E-mail [ M]. Athens, Greece: " DEMOKRITOS", National Center for Scientific Research, 2004.
  • 4METSIS V, ANDROUTSOPIULOS I, PALIOURAS G. Spam filte- ring with naive Bayes -- which naive Bayes? [ C]// CEAS 2006: Proceedings of the Third Conference on Email and Anti-Spam. Mountain View, California: CEAS, 2006:27 -28.
  • 5ZHANG L, ZHU J, YAO T. An evaluation of statistical spam filte- ring techniques [ J]. ACM Transactions on Asian Language Informa- tion Processing, 2004, 3(4): 243-269.
  • 6HOVOLD J. Naive Bayes spam filtering using word-position-based attributes [ C]//CEAS 2005: Proceedings of the Second Conference on Email and Anti-Spam. Palo Alto, CA: CEAS, 2005:41 -48.
  • 7FUMERA G, PILLAI I, ROLI F. Spam filtering based on the analysis of text information embedded into images [ J]. The Journal of Machine Learning Research, 2006, 7:2699 - 2720.
  • 8ALMEIDA T A, ALMEIDA J, YAMAKAMI A. Spam filte- ring: how the dimensionality reduction affects the accuracy of Naive Bayes classifiers [ J]. Journal of Internet Services and Applications, 2011, 1(3): 183-200.
  • 9ALMEIDA T A, YAMAKAMI A, ALMEIDA J. Evaluation of approaches for dimensionality reduction applied with Na- ive Bayes anti-spam filters [ C]// ICMLA'09: Proceedings of the 2009 International Conference on Machine Learning and Applications. Piscataway: IEEE, 2009:517-522.
  • 10HINTON G E, OSINDERO S, TEH Y-W. A fast learning algo- rithm for deep belief nets [ J]. Neural Computation, 2006, 18(7) : 1527 - 1554.

共引文献18

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部