Application of laminated recursive RBF network in loss of customers

• Typical applications • Previous Articles Next Articles

Application of laminated recursive RBF network in loss of customers

Received:2008-05-29 Revised:2008-07-24 Online:2008-11-01 Published:2008-11-01
Contact: Hua-qiu WANG

叠层递归径向基网络在客户流失分析中的应用

王华秋邹航阎河

重庆工学院计算机学院重庆工学院计算机学院重庆工学院计算机学院

通讯作者: 王华秋

Abstract

Abstract: In this paper the laminated recursive Radius Basis Function (RBF) network was applied to the loss of customers, and the solutions of the main issues involved in the loss of customers were presented to address the loss forecast model of the problem, including customer description attributes statute, the reasons for the loss prediction and control strategies. The unit laminated technology was adopted into the recursive RBF network to achieve the dynamic storage, which made the network have a dynamic data processing capability. The self-adaptive distance to optimize the kernel function bandwidth was also presented. The simulation on the temporal data from Mackey-Glass proves the effectiveness of the algorithm. Finally, the trained model was used to predict each customer's loss of possibility and generate the loss list of customers. Through the application to actual cases, the effectiveness and feasibility of the model are verified.

Key words: laminate, recursive Radius Basis Function (RBF), self-adapted distance, loss of customers, prediction model

摘要： 将叠层递归径向基(RBF)网络应用于客户流失研究，给出了客户流失研究中涉及的主要问题的解决方案，包括客户描述、属性规约、流失原因分析以及流失预测与控制策略等，解决了流失预测模型的建立问题，并采用单元叠层输入技术实现了神经网络的动态存储，这使网络具有了动态地处理数据的能力, 提出自适应距离优化核函数的中心半径参数，对混沌时序数据的仿真实验证明算法的有效性。最后利用训练后的模型对每个客户的流失可能性进行预测以及生成可能流失的客户列表，通过对实际案例的应用，对模型的有效性和可操作性进行了验证，证明其具有实际意义。

关键词: 叠层, 递归径向基网络, 自适应距离, 客户流失, 预测模型

Hua-qiu WANG Hang ZOU He YAN. Application of laminated recursive RBF network in loss of customers[J]. Journal of Computer Applications.

王华秋邹航阎河. 叠层递归径向基网络在客户流失分析中的应用[J]. 计算机应用.

[1]	XU Guobao, CHEN Yuanxiao, WANG Ji. Drug-target association prediction algorithm based on graph convolutional network [J]. Journal of Computer Applications, 2021, 41(5): 1522-1526.
[2]	LIU Juan, HUANG Xixia, LIU Xiaoli. Icing prediction of wind turbine blade based on stacked auto-encoder network [J]. Journal of Computer Applications, 2019, 39(5): 1547-1550.
[3]	ZHENG Zhuoran, ZHENG Xiangwei, TIAN Jie. Energy-saving method for wireless body area network based on synchronous prediction with penalty error matrix [J]. Journal of Computer Applications, 2019, 39(2): 513-517.
[4]	SHANG Jiandong, LI Panle, LIU Runjie, LI Runchuan. Recommendation model of taxi passenger-finding locations based on weighted non-homogeneous Poisson model [J]. Journal of Computer Applications, 2018, 38(4): 923-927.
[5]	CHEN Yu, JIANG Wei, ZHOU Ji'en. Compression method for trajectory data based on prediction model [J]. Journal of Computer Applications, 2018, 38(1): 171-175.
[6]	SHI Feng, LOU Wengao, ZHANG Bo. Neural network model for PM_2.5 concentration prediction by grey wolf optimizer algorithm [J]. Journal of Computer Applications, 2017, 37(10): 2854-2860.
[7]	JI Duo, MA Bin, YE Na. Prediction technology of translation query behavior in interactive machine translation [J]. Journal of Computer Applications, 2015, 35(4): 1009-1012.
[8]	LAO Jiajun, YANG Jiang, ZHU Wuming. Dynamic obstacle avoidance method of mobile online water quality monitoring platform [J]. Journal of Computer Applications, 2015, 35(12): 3581-3585.
[9]	ZHANG Yu-mei BAI Shu-lin. Short time traffic flow prediction based on Volterra model using multiplication-coupled configuration [J]. Journal of Computer Applications, 2012, 32(03): 843-846.
[10]	. Target localization scheme for cluster-based beaconless WSN [J]. Journal of Computer Applications, 2007, 27(8): 1835-1838.
[11]	. Compression and application of transition probability matrix in Markov prediction model [J]. Journal of Computer Applications, 2007, 27(11): 2746-2749.