Location-Based O2O Service Recommendation System

The development of mobile Internet technology and the popularity of mobile smart terminals have provided a good technical foundation and hardware conditions for the development of online-to-offline service.The online-to-offline service's market is growing,and a large number of online-to-offline integrated service platforms and vertical applications provide consumers with new consumption choices,meanwhile give a certain amount of advertising bombardment to them,who are difficult to choose from the vast amount of information the most satisfied or worthy of surprise service,especially when the mobile terminal display range is limited and the operating interaction is not as convenient as the PC side.Hence,it requires online-tooffline service to fully consider mobile application scenarios and provide consumers with more effective recommendation lists.The paper aims to provide personalized recommendations for mobile consumers as the starting point,and aims to improve the accuracy of recommendations and users' satisfaction,as well as further analyzes the application context of mobile online-to-offline services,which is realized by the design of the online-to-offline service recommendation algorithm and the application system based on LBS.The paper is based on the online-to-offline service application.It summarizes the research status at home and abroad,studies the advantages and disadvantages of different recommendation algorithms,analyzes the principles of collaborative filtering algorithms and other key theoretical techniques.As the traditional "user-item" twodimensional recommendation model has the disadvantage of collecting the relatively small amount of information,it deeply digs out the online-to-offline service scene information.Besides the basic user information,the paper also collects location,time,weather,user mood and other scene information.And it does the scene information storage,classification,modeling and analysis,in particular with deeply research to the influence of time effect and location information to the recommendation algorithm.After comprehensively measuring the characteristics of multiple similarity calculation methods,it proposes a new multidimensional context collaborative filtering algorithm,and uses Yelp(2017)data to test and verify the precision of the algorithm.Upon this,it further designs and realizes the LBSbased online-to-offline service recommendation system.Through the above research,the research theory of personalized recommendation algorithm is enriched,which provides a scientific basis and an engineering practice reference for the personalized recommendation research of online-to-offline service.