查询结果:   徐涛,刘泽君,卢敏.基于RBM-BPNN的民航潜在高价值旅客预测[J].计算机应用与软件,2019,36(9):58 - 63.
中文标题
基于RBM-BPNN的民航潜在高价值旅客预测
发表栏目
数据工程
摘要点击数
514
英文标题
POTENTIAL HIGH VALUE CIVIL AVIATION PASSENGER PREDICTION BASED ON RBM-BPNN
作 者
徐涛 刘泽君 卢敏 Xu Tao Liu Zejun Lu Min
作者单位
中国民航大学计算机科学与技术学院 天津 300300 中国民航信息技术科研基地 天津 300300 民航旅客服务智能化应用技术重点实验室 北京 101318   
英文单位
College of Computer Science and Technology, Civil Aviation University of China, Tianjin 300300, China Information Technology Research Base of Civil Aviation Administration of China, Tianjin 300300, China Key Laboratory of Intelligent Passenger Service of Civil Aviation, CAAC, Beijing 101318, China   
关键词
民航潜在高价值旅客 特征提取 分类预测模型 RBM BPNN
Keywords
Potential high value civil aviation passenger Feature extraction Classified prediction model RBMBPNN
基金项目
国家自然科学基金项目(61502499);中国民航大学科研基金项目(2013QD18X);民航旅客服务智能化应用技术重点实验室项目(TS-CAKL-2018-01)
作者资料
徐涛,教授,主研领域:民航信息系统理论,智能信息处理。刘泽君,硕士生。卢敏,教授。 。
文章摘要
目前常用潜在客户发现方法多为基于统计特征的行为分析方法,这种方法对所提取的特征具有很强的依赖性并且容易受到人为主观性影响。针对这一问题,结合受限玻尔兹曼机(Restricted Boltzmann Machine, RBM)与BP神经网络(Back Propagation Neural Network, BPNN),提出基于RBM-BPNN的民航潜在高价值旅客发现方法。设置民航旅客类别标签;利用RBM自动提取旅客行为特征;利用BPNN对旅客未来价值类型进行分类预测,从而发现民航潜在高价值旅客。实验结果表明,相对于基于统计特征的行为分析方法,该方法具有更高的分类预测准确率和民航潜在高价值旅客预测效果。
Abstract
Currently, the most commonly used methods of potential customer discovery are the behavior analysis methods based on statistical features, which have a strong dependence on the extracted features and are easily affected by subjectivity. To solve the problem, combining Restricted Boltzmann Machine (RBM) and Back Propagation Neural Network (BPNN), we proposed a potential high value civil aviation passenger prediction method based on RBM-BPNN. The label of civil aviation passenger category was set up; the RBM was used to automatically extract passengers behavior; the BPNN was used to classify and predict the future value types of passengers and to find potential high value civil aviation passenger. The experimental results show that the proposed method has a better classification accuracy and potential high value civil aviation passenger prediction effect than the behavioral analysis methods based on statistical features.
下载PDF全文