Point-of-Interest Recommended:Based On Local Movement Trajectories And Similar Relationships

With the development of location technology and smart devices,traditional social networks have gradually merged with geographic location services to form a Location-Based Social Network(LBSN).Among them,interest points recommendation plays an increasingly important role in location-based social services.Previous research work tends to capture geographic influence from the distance between any two interest points visited by the same user,ignoring two key facts:(1)users are usually in a small area(we call this(2)all points of interest visited by the user are not equally important,some points of interest are repeated multiple times,and some are only visited once.This article is dedicated to portraying users from geographic influences and similar friends and applying them to Point-of-Interest recommendation algorithms.In this paper,our main research contents,contributions and innovations are as follows:1)The local exploratory nature of the user's check-in.In LBSN,most users usually work or live in only one city.That is to say,they mainly generate check-in behavior and explore new locations in a certain city.For this kind of check-in geographical feature,this paper proposes to consider the user's fine-grained time and space attendance rules,mining the impact of geographical impact on the user's check-in,and apply it to the recommended strategy.2)Point-of-Interest recommendations based on local trajectory.As the First Law of Geography says,“Everything is related to everything else,but near things are more related than distant things." For LBSNs,the First Law of Geography means that users like to visit nearby locations rather than distant locations,and locations where the user may be interested in surround the user's favorite location.The user's local trajectory shows the basic rules of the user's movement and special preferences for certain locations,which is more conducive to capturing the user's geographical influence in the area.Based on the user's local movement trajectory,we propose a local trajectory movement model(LTMM)and a corresponding recommendation algorithm.3)Point-of-Interest recommendations based on local similarity relationship.The closer the two users' local activity areas are,the more they may visit the same location.Therefore,we propose local activity similarity(LAS)to calculate the similarity relationship between users.It mainly considers the similarity of the user's local activity area and the cosine similarity between users.Location-based social networks contain rich multi-source heterogeneous information.By mining these information,the performance of POI recommendation system can be improved.Based on the geographical influence of the user's local trajectory and the user's interest similarity relationship,we established a POI recommendation fusion framework LST.The framework integrates the LTMM algorithm and the LAS algorithm,giving full play to the advantages of both.