Research On Recommendation Algorithm Based On Context Awareness And Social Network

With the rapid development of Internet technology,the types and coverage of network data are experiencing a massive growth trend,and the circulation data of various resources is increasing.In addition,users'needs are becoming more diverse and personalized.Many users spend a lot of time and energy searching for resources,but often do not bring the results they want.This causes a waste of resources and affects the user's learning and living experience.In order to solve the problems caused by the redundancy of a large amount of information and the uncertainty of demand,a resource recommendation model based on context awareness and social network is proposed.At present,the most commonly used recommendation algorithm is collaborative filtering,but the traditional collaborative filtering technology only considers the user's historical behavior,and recommends according to the similarity of the scores,and has no connection with the content of the item,the user's own change,the external environment.and the like.Although it has achieved great success in the field of e-commerce,it ignores some unique attributes of users,the relationship between resources,and the relationship between users.It is not suitable for the recommendation of resources in various fields and needs to be improved.After introducing the related theories and advantages and disadvantages of traditional collaborative filtering,this paper first proposes a context-based collaborative filtering algorithm,which uses context entropy and context weights to represent the proportion of different situational factors to resource prediction scores.Then,considering the influence of the social network around the user on the user's choice,if it is a friend's recommended item,the user will be more convinced of the recommendation result,and the higher the affinity with the friend,the higher the degree of user acceptance,so The neighborhood recommendation is weighted with the traditional similarity,and the contextual perception is used to obtain the final recommendation model.Finally,the authoritative real Movielens-1m data set is selected to experiment with traditional collaborative filtering,context-aware collaborative filtering,contextual awareness and collaborative filtering of social networks.The results verify the recommendation of the contextual awareness and social network recommendation model.The result is superior to traditional collaborative filtering.In this paper,according to the actual needs of different users,the recommended resources are effectively recommended to the users,and the services of the recommendation system are passive and active,and the optimization of the recommended services is realized,which not only improves the utilization of resources,but also saves users' time and energy.