Research And Application Of Recommendation Algorithm For Integrating Time And Space Information And User Trust

The recommendation algorithm based on location service is mainly based on the user's check-in data,analyzing the relationship between the geographic location characteristics of user's activities and user's social activities,and establishing the location recommendation model.Traditional location-based recommendation systems mostly use user's context information,such as user's location and user's registration information,to recommend users,but ignore that user's interest will change with time or location,and trust between users and friends is also an important factor affecting user preferences.Aiming at the migration of user preferences,this paper mine the user's check-in data sets,and use the information of user's check-in time,check-in place,user rating data,user's social situation and so on to model the user.The main work is as follows.Firstly,according to Newton's cooling law,a time attenuation function is proposed.In space,a spatial similarity calculation method is proposed based on the longitude and latitude information of the user's location.Then,the space-time information based on the two methods is integrated into the collaborative filtering algorithm,and the proposed recommendation algorithm is synthesized.The algorithm is compared and analyzed with other mainstream recommendation algorithms.The experimental results show that the proposed algorithm is superior to other comparative recommendation algorithms.Secondly,in social networks,the trust of users' friends greatly affects users' preferences.Therefore,this paper uses the number and length of communication in the mobile phone communication records of mobile users to establish a special mobile user friend-neighbor relationship model.At the same time,considering the different communication modes of users,set user weight,and calculate the trust degree of users to direct and indirect friends,forming a set of users' nearest neighbors,and generate top-N recommendation list.Experiments on data sets show that the proposed algorithm is superior to other comparison algorithms in alleviating data sparsity and cold start of recommendation system.Finally,based on the above algorithm research,this paper designs and implements a personalized food recommendation system,which makes full use of the implicit information of the data and satisfies the user's demand for personalized food recommendation,having a certain practical reference significance.