Research And Implementation Of POI Recommendation Algorithm Based On Spatiotemporal Information And Social Network

Recently,as location-based social network services(LBSNs)become more popular in human's lives,POI(Point-of-Interest)recommendation has become an important research topic.Its purpose is to recommend places that the user has not visited before but may be interested in,and has a high value to users and POI owners.In this paper,we study the POI recommendation problems based on temporal information,geographic information and social network,mainly include the following aspects:Firstly,we study the POI recommendation algorithm based on social network and geographic information,and propose a model that called SROF.In the mode,we consider the different types of user relationships,such as user similarity,social relationship and neighbor relationship.At the same time,we split the business district by the geographical information of different POI and calculate the probability of the users from one POI to another.Experiments show that this model improves the accuracy of the recommendation.Secondly,we study the POI recommendation algorithm based on temporal information and geographic information,first propose the CTS model.In the mode,we exploit the user check-in data to compute temporal behavior similarity between users,and use it to supplement the data to solve the problem of data sparse.At the same time,according to the user's check-in records to split the user's reachable distance,the candidate points are selected and the popularity of different POIs in different time periods is combined.Moreover,we propose the S-CTS model what combined with the previous content.Experiments show that CTS model and S-CTS model has a good effect in real-time POI recommendation.Newly,we study the POI recommendation algorithm based on POI category and geographic information,multi-category locations were processed and propose the GUP model what combine the matrix factorization technique.In the model,we exploit the user check-in records and POI category to get the user preference matrix by the matrix factorization method,and combine it with geographic information for POI recommendation.Moreover,the ST-GUP model is proposed in combination with the above conditions.Experiments show that GUP model and ST-GUP model can improve the accuracy of the recommendation in a certain extent.Finally,we compare the performance for the five methods on the real data sets.Experimental results show that the performance of S-CTS model and ST-GUP model are better than other methods,it shows that combine multiple types of data information together is beneficial to improve the performance of POI recommendation,and the ST-GUP model is superior to the S-CTS model.