Research And Implementation Of A Personalized Location Recommendation Algorithm Based On User Trajectory

With the development of China economy,the mobile Internet has been widely popularized.Communication between people is becoming more and more frequent.In these communication activities,a great deal of user movement track information is generated,from which we can know users' travel habits,consumer preferences and the popularity of the business circle.Currently,the location-based recommendation system that is widely used in the community is mainly based on the current user's distance from nearby shops and does not take into account the user's consume habit and regional heat,making the recommendation results not personalized and reasonable.This paper analyzes the relevant personalized shop recommendation system,proposes a new improved algorithm based on the user's movement track and forms a recommendation system.The main contents of this paper are three aspects:(1)In this paper,Spark-based big data parallel computing framework and Neo4 j network graph are researched,which help us analyzes and preprocesses massive mobile trajectory data.(2)Based on the research of traditional K-Means algorithm,this paper proposes an improved K-Means algorithm based on Spark computing platform.Experiments show that the improved algorithm is more efficient than the original one in terms of iterative efficiency and running time,and shows better speedup and scalability when dealing with large-scale data.(3)This paper designs and implements a personalized business recommendation system based on user's trajectory by introducing the concept of path network.Experiments show that the recommendation system has high recommendation accuracy and effectiveness.