Research On Location Recommendation Technology Based On Social Relationship

As the Mobile network,social media,and mobile location technology have made rapid development in recent years,users can easily access their current location and thus,location-based social networks(LBSNs)are also born.Users in LBSNs can use mobile devices such as mobile phones to check-in at the currently visited location and share with friends.However,with the rapid increase in the number of users and the number of sites,how to mine the preferences of users in the massive data information and provide personalized POI recommendation for users has become a very popular research direction in the field of recommendation.A large number of scholars and experts have put forward many valuable POI recommendation algorithms,but due to the influence of cold start,data sparsity,and other problems,the existing POI recommendation algorithms based on LBSNs are not as good as expected.Based on the advantages and disadvantages of some existing classic POI recommendation algorithms and the characteristics of the location social network,this paper uses information such as social relationships to help predict users' interest preferences and improve the accuracy of personalized location recommendations.The following are the main research results of this article:1)Neumf model based on neural matrix decomposition is a location recommendation model based on deep learning.It is a combination of GMF and MLP models based on generalized matrix decomposition.2)This paper improves the GMF and MLP models,respectively,by integrating social relations into GMF and MLP models.3)Experiments on two open datasets,Brightkite and Gowalla,show that the two improved models are better than the original model,and the improved Neumf model is also better than the original model.Some other classical algorithms based on matrix decomposition algorithm are also used for comparison and the experimental results show that the improved Neumf is better than these algorithms.