Research On The Design And Application Of Railway Freight Key Customer Portrait Based On Big Data

In China,railway freight transportation bears the medium and long distance transport missions of materials needed for social development and economy and the material transportation task of strategic significance.It is an important guarantee of material transport and a key component of logistics system in China.With the rapid development of social economy and the gradual improvement of people’s living standards,the demand for production and living materials increases,and then the demand for cargo transportation increases.As a result,new modes of cargo transportation emerge as the times requiring,and the market competition in the freight transportation industry becomes increasingly fierce.In the era of both opportunities and challenges,railway freight transport has not continued its advantages.On the contrary,it is at a disadvantage in the market competition,and the loss of customers is serious.Therefore,how to effectively tap customers is the primary problem that railway enterprises should solve if they want to gain a firm foothold and move forward.At present,the amount of research data for big customers is generally small,and the data source is single.Therefore,this paper integrates internal and external data on the basis of big data to analyze railway freight customers.A key customer model L(Length)R(Recency)F(Frequency)A(Average Monetary)V(Rotation Volume of Freight Transport)is proposed,and A key customer individual portrait model KCPM is established.Finally,on the basis of the individual portrait of the key customers,the loss of the railway freight customers is predicted.The specific research work is as follows:(1)Combining the characteristics of the industry and the existing data characteristics of railway freight customers,the freight key customer model LRFAV based on RFM model is proposed by cleaning data,protocoling data and transforming data.In other words,on the basis of RFM,two parameters,relative passenger age of freight customers(L)and freight turnover of freight customers(V),are introduced,and M is converted into A.The comparison between the RFM model and the LRFM model is made with the real railway freight ticket data to verify the good customer segmentation effect of the LRFAV model.At the same time,under the guidance of senior experts in the railway freight industry,AHP method was used to assign corresponding weights to the five variables of the model.Then,under the guidance of experts,AHP method is used to assign corresponding weights to model variables of key customers,and the classification of railway freight customers is combined with K-means ++ clustering algorithm to calculate the customer value of all kinds of customers,and finally select the key customers.(2)Firstly,the feasibility of the construction of key customer portrait model is analyzed.Next,the selected key customer invoice data and non-key customer invoice data are randomly mixed at a rate of 1:4 to form new data.Climb the external data of customers in the new data on the company’s official website,and fuse it to obtain the portrait data of big customers.Then extract the analysis characteristic index of the customer portrait from the key customer portrait data,and carry out data preprocessing,chi-square independence test,Point Biserialr(point-two-line correlation coefficient)test,Pearson correlation coefficient test,and collinearity diagnosis to remove the redundant variables to obtain the analysis data set of key customer portrait.Finally,K-Means clustering is used to segment the values of variables in the analysis data set of key customer portrait,and the key customer individual portrait model KCPM is obtained.(3)Analyze data characteristics based on key customer profiles and define churn customers.Logistic regression,decision tree and BP neural network are respectively used to predict the loss of key customers in railway freight transport.The comparison shows that the prediction effect of BP neural network method is better,and the recall rate of railway freight loss major customers reaches 95%,which basically meets the forecast demand.This is conducive to the railway group to retain freight big customers,reduce the occurrence of heavy losses.