Design And Implementation Of Car Rental Recommendation System Based On Microservice Architecture

With the improvement of people’s income level and the change of consumption concept,the car rental industry has been rapid development.With so many car brands and car types,people can’t pick out the car they are interested in in the time available.Recommendation systems can help users quickly find the information they are interested in under the condition of information overload.In order to help users to quickly select the car they are interested in when the target is not clear,the system also introduces personalized recommendation function and similar recommendation function.This thesis proposes a car rental recommendation system based on microservice architecture.The microservice architecture is used to build the entire system framework to ensure the stability and expansibility of the system.Meanwhile,the deep learning recommendation algorithm is used to complete personalized recommendation functions and similar recommendation functions to ensure the accuracy of car recommendation.The main work contents are as follows:（1）in this thesis,based on factorization machine layer and multilayer perceptron sharing embedded deep learning recommendation model,achieve the personalized recommendation feature of car rental recommendation system,factorization machine has good characteristics of crossing ability and memory ability and multilayer perception machine has the strong ability of fitting,a combination of the data mining in car rental recommendation model.（2）This thesis uses the Deep Walk algorithm based on random walk to generate embedding vector.Deep Walk algorithm can not only process sequence data,but also data in the form of graph structure.More association information between users can be mined through random walk,which solves the problem of insufficient data set in the training process of recommendation model.（3）Based on the realization of personalized recommendation by users,this thesis adds the similar recommendation function into the system,realizing the car rental recommendation system combining personalized recommendation by users and similar recommendation by cars.This thesis proposes and designs a method based on embedding vector to realize the similar recommendation function of cars and help users to conveniently select the cars they are interested in.（4）In this thesis,the combination of microservice architecture and deep learning recommendation algorithm is proposed to realize the car rental recommendation system.Compared with single application,microservice has higher stability and expansibility.The system builds the microservice architecture through the combination of Spring Cloud and Spring Cloud Alibaba,which reduces the coupling degree of system components.Meanwhile,it designs and implements the management functions of cars,orders,evaluation and users,providing stable background support for the recommendation function.（5）The system crawls the vehicle data set and preprocesses the data set using Spark platform.Tensor Flow is used to train,evaluate and deploy the recommendation model.Finally,matplotlib is used to print the training results.Based on the above work content,this thesis designed the car rental recommendation system based on microservice.After the training evaluation and system test of the recommendation model,the experimental results are stable and meet the expected goals.