自组织人工神经网络在民航旅客盈利能力细分中的应用

Application of artificial neural network in airline customer segmentation

School of Civil Aviation, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China

摘要: 目前采用的基于里程的旅客划分方法不能清晰区分旅客价值, 应用Kohonen T自组织特征映射的人工神经网络模型将反映旅客盈利能力的多维行为特征属性数据以有序的方式映射到对旅客盈利性判别等级的低维空间, 形成对旅客正确识别的一种拓扑意义的有序图。试验结果表明, 此模型对客户细分识别成功率较高, 准确率可达90.8%。
- 民航运输 /
- 客户细分 /
- 人工神经网络 /
- KohonenT自组织模型
Abstract: The segmentation of airlines passengers based on mileage can not reflect passengers characteristics. The model that can show passengers value was developed. In this model an artificial neural network was applied to give a topological maps, with the maps, the data that can indicate passengers value was collected. An example proves that the accuracy to seperate passengers by their value with this model is 90. 8%.
- civil aviation transportation /
- passengers'segmentation /
- artificial neural network /
- Kohonen T self-organizing model

图 1 SOM网络结构

Figure 1. Layout structure of SOM

下载: 全尺寸图片幻灯片

图 2 训练样本的SOM二维映射

Figure 2. SOM of training sample

下载: 全尺寸图片幻灯片

图 3 SOM的二维U矩阵

Figure 3. SOM 2-D U-matrix

下载: 全尺寸图片幻灯片

图 4 SOM的三维U矩阵

Figure 4. SOM 3-D U-matrix

下载: 全尺寸图片幻灯片

表 1 旅客盈利能力分类

Table 1. Classification of passengers'profitable ability

下载: 导出CSV

表 2 部分学习样本数据

Table 2. Training sample datum

下载: 导出CSV

表 3 验证样本数据

Table 3. Testing sample datum

下载: 导出CSV

[1]	田静. 常客优惠筑就忠诚[J]. 中国民用航空, 2001, 1(6): 29-31. https://www.cnki.com.cn/Article/CJFDTOTAL-MHJJ200106010.htm TIAN Jing. Preferential treatment to frequent flyer builds loy-alty[J]. China Civil Aviation, 2001, 1(6): 29-31. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-MHJJ200106010.htm
[2]	黄德双. 神经网络模式识别系统理论[M]. 北京: 电子工业出版社, 1996.
[3]	Simon Haykin. Neural Networks: A Comprehensive Founda-tion[M]. 北京: 清华大学出版社, 2001.
[4]	Kohonen T. Self-Organization and Associative Memory[M]. New York: Springer-Verlag Berlin Heidelberg, 1989.
[5]	Berry M, Lindoff G. Data Mining Techniques for Marketing, Sale and Customer Support[M]. John Wiley and Sons, 1997.
[6]	飞思科技产品研发中心. Matlab 6. 5辅助神经网络分析与设计[M]. 北京: 电子工业出版社, 2003.
[7]	Kohonen T. Self-Organizing Map[M]. Berlin: Springer-Verlag, 1995.
[8]	Siemon H P. Kohonen's self-organizing feature maps for ex-ploratory data analysis[A]. INNC'90, International Neural Network Conference[C]. Dordrecht, Netherlands, 1990.