Research On Passengers' Fancy-oriented Recommendation Algorithm Of Online Ride-hailing Service

The rapid development of the internet and mobile internet has created the big data era and sharing economy era.And it is hard for people to take a taxi in rush hours or in remote areas.All these reasons make the application of online ride-hailing service come into being.In order to ensure the user's loyalty to the taxi application,the key method is to provide users with high-quality personalized service.Recommender system is the technology of providing personalized service.Therefore,it is meaningful to research the recommendation algorithm of online ride-hailing service in order to provide accurate and efficient recommendations.This paper collects the information of the development in taxi application,and analyzes the current status,then puts forward the goal of designing personalized driver recommendation algorithm to meet passengers' fency.First of all,analyze the current research review of recommendation algorithm on online ride-hailing service at home and abroad.It has found a large number of research production about recommendation algorithm in various fields,but has not yet found there is a study of the application of the recommendation system in the field of reserving a taxi online.So the research in this paper is valuable.Confirm the recommendation algorithm in the model by presenting the related theories about the recommendation system,clustering,context awareness and other correlative knowledge.Secondly,establish the model of recommendation.The model building process includes three parts:No.1 Establish collaborative filtering recommendation model.According to the passenger's history rates to driver,the traditional item based collaborative filtering algorithm is adopted to recommend for the target users.No.2 Establish speed improvement model based on Clustering.Fuzzy clustering algorithm is introduced to improve the efficiency of recommendation.Cluster passengers and drivers according to the attribute information of passengers and drivers respectively,and then recommend by collaborative filtering recommendation algorithm.No.3 Establish precision improvement model based on context aware.Incorporate the concept of context-aware into recommendation algorithms by Byes algorithm to improve the accuracy of recommendation.Make combinatorial model with context information,rates information,passenger attribute and driver attribute.Then recommend by collaborative filtering recommendation algorithm.Finally,select the experimental platform,access to relevant experimental data,and preprocess data according to the experimental requirements.Verifie the three models on the obtained data set.The experimental results showed that the proposed recommendation algorithm based on clustering and context awareness improves thespeed and accuracy of recommendation,and the algorithm satisfies the requirement of recommendation in this system.