Research On Predicting The Loyalty In Passengers Of Civil Aviation

In recent years,with the rapid growth of the quantity of airlines in China,especially low-cost airlines,the competition between airlines has become increasingly fierce,making the loyalty of civil aviation passengers more difficult to maintain.Therefore,it is urgent to increase and retain the loyalty of civil aviation passengers.At the same time,civil aviation informationization is accelerating,and civil aviation customer relationship management is steadily moving towards the direction of data.it is important research significance how to use passenger travel records to accurately predict passenger loyalty of civil aviation for airlines to explore potential loyal travelers.This article firstly defined the concept of civil aviation passenger loyalty and elaborates the characteristics of civil aviation passenger loyalty.This paper discussed the inadequacy of the traditional questionnaire survey method to study the loyalty of civil aviation passengers,and determines the objective data used by civil aviation Passengers Name Record to evaluate civil aviation passenger loyalty.The civil aviation passenger loyalty evaluation system was built with seven indexes including the number of passengers’ historical ticket purchases,flight mileages,recent purchase time,length of customer relationship,average discount,number of group purchases,and number of round trips,and the passengers’ loyalty was analyzed using AHP.Degree evaluation indicators was empowered.The data of travel records of 20,000 passengers of A Airlines was extracted and the passenger loyalty evaluation values were calculated.Using the K-means cluster analysis method,the assessed passenger sample was divided into high,medium and low categories according to the loyalty of passengers.Finally,from the civil aviation passengers analyzed by K-means clustering,there are 12,916 number of passengers with consumption records in 2017 who were extracted as samples and divided into training set and test set with a ratio of 2:1.The training set was used to construct the Bayesian network model.The behavioral attributes of passengers in one year were extracted as the nodes of the Bayesian network.Bayesian network was studied structurally by the independence test method and the passengers of Bayesian Airlines were obtained.Network structure diagram;learning from the maximum likelihood estimation,obtaining the conditional probability table of each passenger behavior attribute;predicting the category of passenger loyalty through the maximum posterior probability method;and using the test set to predict the Bayesian network The ability was evaluated and the Bayesian network was better at predicting high loyalty passengers.