Research On Personalized Service Recommendation Model For Expressway ETC Users Based On Portrait Analysis

Since the expressway ETC national networking,ETC has realized the business transformation from traditional toll collection to multi-scenario integration development.Expanding ETC multi-scene services and realizing the industrial integration of ’ETC +Internet’ have become the future development trend.At the same time,with the continuous improvement of people’s living standards and the rise of the concept of Maa S,people’s demand for travel quality and personalization is also increasing.Exploring the characteristics of user travel and consumption,using ETC multi-scene services to meet people’s travel needs,and providing users with comfortable and convenient personalized travel services will become the focus of future research.Based on the multi-source data of expressway ETC,this thesis analyzes the characteristics of users’ travel and consumption,constructs a portrait of ETC users’ travel and consumption characteristics,and combines it with the recommendation algorithm to propose a Personalized Service Hybrid Recommendation Algorithm Combined with User Portrait,which realizes the personalized recommendation of users’ travel services.The main contents are as follows:(1)Through the multi-source data such as toll collection,user attributes,vehicle information and user consumption transactions accumulated by ETC business system for a long time,an ETC user travel and consumption feature portrait model with three levels(data layer,label layer,application layer)and five dimensions(personal attributes,user travel,consumption attributes,user preferences,user sensitivity)is constructed.Among them,the data layer implements data preprocessing such as cleaning,processing,expansion,and association of multi-source data;the label layer uses statistical analysis,K-means clustering,Term Frequency-Inverse Document Frequency(TF-IDF)and other methods to propose a set of ETC user travel and consumption feature portrait label system.The application layer uses word cloud analysis,clustering algorithm and other technologies to realize the visual display of ETC users from individual dimension and group dimension.Through the construction of ETC user travel and consumption feature portraits,it lays a foundation for the construction of user personalized recommendation algorithms.(2)A Personalized Service Hybrid Recommendation Algorithm Combined with User Portrait(UPPRA)is proposed to solve the problem of user personalized recommendation and data sparsity.Aiming at the problem of personalized recommendation,A Recommendation Algorithm Based on User Space Vector(USVRA)is designed to realize travel service recommendation based on user history data and preference records.Aiming at the problem of sparse data in users’ historical consumption records,the user portrait method is introduced to convert the user-item rating matrix in the traditional recommendation algorithm into the user-user portrait preference matrix,and then A Recommendation Algorithm Based on User Portrait(UPRA)is proposed.Finally,the linear weighting method is used to weight The Recommendation Algorithm Based on User Space Vector(USVRA)and The Recommendation Algorithm Based on User Portrait(UPRA),and A Personalized Service Hybrid Recommendation Algorithm Combined with User Portrait(UPPRA)is proposed.(3)The proposed Personalized Service Hybrid Recommendation Algorithm Combined with User Portrait(UPPRA)is compared with the traditional single recommendation algorithm.The results show that UPPRA is better than other algorithms in MAE index,Precision index and Recall index.It alleviates the problem of data sparsity while satisfying the personalized recommendation problem,and has strong practical application value for improving user travel and consumption experience.