Study On Customer Value Of Airline Company Based On Multivariate Statistical Method

With the large number of Chinese airlines and the rapid development of high-speed rail,airlines are facing a huge competitive pressure.However,the competition between airlines is essentially the competition of customer resources and customer value.Therefore,in order to develop continuously and healthily in a competitive society,the airline must fully understand the needs of customers and provide the most appropriate service according to the needs of customers.How to understand the different needs of different customers,how to meet the needs of most customers with the minimum cost.It is an effective solution that using the very popular data mining tools in multivariate statistics,for airlines to massive amounts of customer information in the database,data mining according to the result of customer segmentation for differences between different levels of customers with personalized service.This paper takes southern airlines as an example to focus on this solution.Firstly,this paper elaborates the customer value theory,the theory of customer relationship management,customer life cycle theory,customer segmentation method,etc.Then,the data mining algorithm in multivariate statistics is introduced in detail.After that,we used Python to deal with the customer data in the southern airlines database,and used k-means clustering method to divide the customers into five levels of unequal value.Finally,the R language as a data mining tool,used the decision tree in data mining,random forests,support vector machine,artificial neural network algorithm is constructed to predict customer level such as model and the future development trend of model prediction and old customers.After the segmentation of customers,the southern airlines can conduct one-to-one personalized service for customers according to the classification resul ts,and realize accurate marketing.In this way,southern airlines can apply the limited marketing resources to the most appropriate places,so that the quality customers can enjoy more comfortable service.For important to retain customers and the general customers,southern airlines also should take active measures to try to convert these customers into important development customer or keep customers,so that China southern airlines can not only save costs,you can also get more high quality sustainable customer resources,to enhance the comprehensive strength of China southern airlines.The model that builds the prediction customer hierarchy is the first innovation point of this article.In order to improve the model of forecasting and old customers to the future development trend,this paper constructs two random forest prediction function to predict whether old customers shall emphatically customers,one of which is a high hit rate function,a function is a high recall rate.Through the combination of the two,the advantages complement each other and greatly improve the prediction effect of the model.This is the major breakthrough in the method of this paper and the second most valuable innovation in this paper.