Research On Identification And Prediction Of Abnormal User Of Carrier Of NTOCC Based On Data Mining

In 2013,the Ministry of Transport introduced the "car-free carrier" model,and issued a number of policies and measures in the following years to standardize and promote the development of the car-free carrier platform,and officially renamed the car-free carrier platform in 2020 It is a network freight platform.The emergence of the network freight platform has solved the problem of high logistics costs in China’s road freight market to a certain extent.With the rapid development of the network freight platform,more and more carriers and shipper users have joined the network freight platform.The network freight platform relies on mobile Internet technology.In recent years,with the rapid development of the mobile Internet,various abnormal users have appeared in the mobile Internet,and the network freight platform with mobile Internet characteristics also has abnormal users.Due to the large scale of users of the network freight platform carrier,which involves cargo transportation and freight delivery and other behavioral operations,this article identifies the research subject as the abnormal user of the network freight platform carrier.During the operation of the network freight platform,a large amount of order transaction data and carrier user data are generated.In order to help the network freight platform to better identify and predict abnormal carrier users,this article hopes to effectively use the network freight platform big data through data mining technology,Identify and predict abnormal users of the platform carrier,thereby providing beneficial support for the network freight platform to scientifically manage carrier users.With the help of data mining technology,this article takes the network freight platform A as an example to study the identification and prediction of abnormal users of the platform carrier.This paper first discusses the research status of the identification and prediction of network freight platforms and abnormal users,and introduces the method theory of network freight platforms,abnormal users,and data mining.Since the abnormal users of network freight platform carriers have not been clearly defined,this article draws on Scholars have given the definition of abnormal users and the definition of abnormal users in various fields to the definitions of abnormal users and normal users of network carriers.Then introduced the data mining related knowledge,and took the network freight platform A as an example,extracted order data and carrier basic data from the platform A,data preprocessing,exploratory analysis and feature engineering operations on the data,based on the indicator system construction Principles and methods,combined with the actual situation of platform A,constructed a carrier abnormal user index system.Finally,standardize the processed data,use K-Means clustering to cluster the data of platform A ’s carrier users and analyze the clustering results,and classify platform A users into normal users,high-risk abnormal users and For low-risk anomalous users,the characteristics of different user groups are analyzed separately,and related countermeasures and suggestions are proposed for A platform to identify abnormal users and manage different user groups.Considering that the network freight platform will predict whether more carriers are abnormal users,the K-Means clustering result is used as the original data set.Since the sample data of the three user groups is in an unbalanced state,the SMOTE oversampling method is used to make the sample size Balanced,using three classification algorithms of random forest,SVM and Xgboost to perform classification prediction,analyze the prediction results and put forward relevant suggestions for platform A to help network freight platform identify and predict abnormal users of carriers.This paper uses data mining technology to identify and predict the network freight platform carrier,and puts forward relevant countermeasures and suggestions based on the results,helps platform management users provide data support,improves platform carrier user stickiness and loyalty,and network freight The development of the platform has high practical significance and reference value.