面向民航旅客同行特征提取与设计  

Feature extraction and design for peer-passengers in civil aviation

作　　者：徐涛[1,2] 邢泽文 卢敏[1,2] 李忠虎[3] XU Tao;XING Ze-wen;LU Min;LI Zhong-hu(College of Computer Science and Technology,Civil Aviation University of China,Tianjin 300300,China;Information Technology Research Base of Civil Aviation Administration of China,Civil Aviation University of China,Tianjin 300300,China;Key Laboratory of Intelligent Passenger Service of Civil Aviation,TravelSky Technology Limited,Beijing 101318,China)

机构地区：[1]中国民航大学计算机科学与技术学院,天津300300 [2]中国民航大学,中国民航信息技术科研基地,天津300300 [3]中国民航信息网络股份有限公司,民航旅客服务智能化应用技术重点实验室,北京101318

出　　处：《计算机工程与设计》2021年第2期589-594,共6页Computer Engineering and Design

基　　金：天津市自然科学基金项目(18JCYBJC85100);教育部人文社会科学研究规划基金项目(19YJA630046)。

摘　　要：为挖掘民航旅客潜在同行关系,构建完善的旅客同行网络,提出从民航旅客订票记录进行民航旅客同行特征提取算法。通过计算信息熵等发现特征相关性,提取旅客同行表现出强相关性的特征,细化设计同行旅客对的特征集合。实验结果表明,各特征均反映了不同强度的旅客同行关系,利用特征向量对基础分类器模型进行训练预测,平均准确率高达0.91,验证了该方法具有极高的适用性。To explore the potential peer relationship of civil aviation passengers and to construct a complete network of passengers,a civil aviation passenger peer feature extraction algorithm from civil aviation passenger booking records was proposed.The correlation of features was found by calculating information entropy,the characteristics of strong correlation between passengers were extracted,and a collection of features for peers was designed.Experimental results show that each feature reflects the passenger relationship of different strengths.The eigenvectors are used to train and predict the basic classifier model,and the ave-rage accuracy is as high as 0.91,which verifies that the feature extraction method has high applicability.

关 键 词：民航旅客同行关系 特征提取 信息熵 特征向量化 分类器模型 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]