针对目前民航旅客关系网络中旅客关系类别单一、旅客价值计算滞后、旅客潜在价值计算没有考虑旅客之间的相互影响，使得一些具有极大消费潜力的旅客因为目前乘机次数较少而被航空公司忽略的现状，提出了一种融合个体属性与社交关系的民航旅客价值度量方法，采用RFMc模型计算旅客个体价值，并采用多关系评价（MRE）模型分析旅客关系，然后通过改进的PageRank算法设计实现融合旅客个体价值和社交关系的民航旅客价值排序（CAPV-Rank）算法，实现旅客价值度量、旅客价值预测和潜在高价值旅客挖掘。实验结果表明：设计实现的CAPV-Rank算法可通过调整权重因子实现多种模式下的旅客价值计算，满足各种业务需求，并能实现旅客价值预测、潜在高价值旅客挖掘，旨在为民航旅客价值度量和预测提供灵活、高效的解决方案。
In the civil aviation passenger network, the type of relations among passengers is single, and calculation of passenger value is lagging behind. The calculation method for the passenger potential value does not take into account the interaction between passengers, making the passengers who seldom travel by plane but have great potential for consumption in the future ignored by the airlines. A method for measuring the civil aviation passenger value by combining individual attributes and social relations is proposed. First, the RFMc model is used to calculate the individual value of passengers, and the MRE model is adopted to analyse the relationship of passengers. PageRank is then improved to implement the proposed Civil Aviation Passenger Value Rank (CAPV-Rank) algorithm, which merges the individual value and social relationship of the passengers. The CAPV-Rank algorithm is designed to realize the passenger value measurement, the passenger value forecast and potential high value passenger mining. Experimental results show that the CAPV-Rank algorithm can implement the passenger value measurement in various modes by adjusting the weighting factor, and can satisfy various business requirements. The algorithm can also realize the passenger value forecast and potential high-value passenger mining, providing a flexible and efficient solution for the measurement and forecast of civil aviation passenger value.
[1] 冯霞, 徐冰宇, 卢敏. 民航旅客订票行为细分及群体特征分析[J]. 计算机工程与设计, 2015, 36(8):2217-2222. FENG X, XU B Y, LU M. Booking behavior subdivision and characteristic analysis of civil aviation passenger[J]. Computer Engineering and Design, 2015, 36(8):2217-2222(in Chinese).[2] 潘玲玲. 基于旅客行为的航空旅客细分模型研究及其实现[D]. 南京:南京航空航天大学, 2012:1-57. PANG L L. The research and realization of civil aviation customer segmentation based on customer behavior[D]. Nanjing:Nanjing University of Aeronautics and Astronautics, 2012:1-57(in Chinese).[3] 林友芳, 王琨琨, 周超, 等. 基于社交网络的民航旅客偏好建模[J]. 北京交通大学学报, 2014, 38(6):33-39 LIN Y F, WANG K K, ZHOU C, et al. Modeling the preference of air passengers based on social network[J]. Journal of Beijing Jiaotong University, 2014, 38(6):33-39(in Chinese).[4] 王坤坤. 民航旅客座位偏好建模与应用研究[D]. 北京:北京交通大学, 2015:1-48. WANG K K. Research of modeling the seat preference of civil aviation passengers and its applications[D]. Beijing:Beijing Jiaotong University, 2015:1-48(in Chinese).[5] 曹卫东, 白亮, 聂笑盈. 基于Map/Reduce的民航高价值旅客发现方法[J]. 计算机工程与设计, 2015, 36(4):1078-1083. CAO W D, BAI L, NIE X Y. Method of discovering high-value passengers of civil aviation based on map/reduce[J]. Computer Engineering and Design, 2015, 36(4):1078-1083(in Chinese).[6] FENG X, XU B Y, MIN L, et al. Potential high-value passengers discovery by random walk on passenger-route heterogeneous network[J]. Journal of Computational & Theoretical Nanoscience, 2015, 12(8):1568-1593.[7] 韩敏. 基于社会网络的民航旅客价值排序算法研究与实现[D]. 北京:北京交通大学, 2014:22-38. HAN M. The research and implementation on ranking the aviation passengers' values based on social network[D]. Beijing:Beijing Jiaotong University, 2014:22-38(in Chinese).[8] 冯霞, 李勇, 陈卉敏. 民航旅客社会网络构建方法研究[J].计算机仿真, 2013, 30(6):51-54, 142. FENG X, LI Y, CHEN H M. Research on constructing social network of airline customers from data of PNR[J]. Computer Simulation, 2013, 30(6):51-54,142(in Chinese).[9] HAVELIWALA T H. Topic-sensitive PageRank[C]//International Conference on World Wide Web, 2002:517-526.[10] KAMVAR S D, HAVELIWALA T H, MANNING C D, et al. Exploiting the block structure of the web for computing PageRank[R]. Palo Alto, San Francisco:Stanford University Technical Report, 2003:1-13.[11] YANG Z, TANG J, ZHANG J, et al. Topic-level random walk through probabilistic model[M]//Advances in Data and Web Management. Berlin:Springer Berlin Heidelberg, 2009:162-173.[12] 朱凡微, 吴明晖, 应晶. 高效个性化PageRank算法综述[J]. 中国科技论文, 2012, 7(1):7-13. ZHU F W, WU M H, YING J. Efficient personalized PageRank computation:A survey[J]. China Sciencepaper, 2012, 7(1):7-13(in Chinese).[13] ZHU F W, FANG Y, CHANG C C, et al. Scheduled approximation for personalized PageRank with utility-based hub selection[J]. The VLDB Journal, 2015, 24(5):1-25.[14] WEI W, GAO B, LIU T Y, et al. A ranking approach on large-scale graph with multidimensional heterogeneous information[J]. IEEE Transactions on Cybernetics, 2016, 46(4):930.[15] WAN X, XIAO J. Single document keyphrase extraction using neighborhood knowledge[C]//National Conference on Artificial Intelligence, 2008:855-860.[16] LI D, LI S, LI W, et al. A semi-supervised key phrase extraction approach:learning from title phrases through a document semantic network[C]//Proceedings of the, Meeting of the Association for Computational Linguistics, 2010:296-300.[17] SIDDIQI S, SHARAN A. Keyword and keyphrase extraction techniques:A literature review[J]. International Journal of Computer Applications, 2015, 109(2):18-23.[18] MIHALCEA R, TARAU P. TextRank:Bringing order into texts[J]. Unt Scholarly Works, 2004:404-411.[19] AMJAD T, DING Y, DAUD A, et al. Topic-based heterogeneous rank[J]. Scientometrics, 2015, 104(1):1-22.[20] DING Y. Topic-based PageRank on author cocitation networks[J]. Journal of the Association for Information Science and Technology, 2011, 62(3):449-466.
CJA