Research On Edge Service Enhancement Strategy Based On Dynamic Interest Capture

The large increase of end-users' demand for multimedia services has led to the exponential growth of mobile data traffic,resulting in data overload transmission on the network.Caching service data on the edge network can greatly reduce the latency of user content requests.However,the limited storage capacity of edge servers makes the cache deployment strategy directly affect the response performance of edge services to end users.At the same time,scientific and reasonable recommendations can increase user demand for edge cache content and improve the quality of edge services.Therefore,from the perspective of end users,it is of great significance to accurately capture users' interests and preferences and design edge Service strategies to improve the Qo E and Qo S of users.An edge cache deployment strategy based on cross-domain knowledge transfer is proposed to solve the problem of unbalanced user content requests and limited edge node cache capacity.The transfer learning technology is used to transfer the relevant knowledge of the source domain model composed of one-dimensional extended CNN to the target domain,and the user preference perception is realized through further learning.The autoregressive model is used to predict the user trajectory and update the service list of edge nodes.An edge cache deployment strategy based on user preference perception is designed so that the edge nodes can respond to the content requests of their service users without relying on the backhaul when the remaining cache capacity is insufficient.Simulation results show that the proposed strategy can significantly improve the cache hit rate and the user's Qo E.Aiming at the problem that users' dynamic characteristics and recommendation system can reduce the probability of users' request for non-recommended content,a content recommendation strategy of edge cache based on dynamic interest capture is proposed.Deep learning technology is applied to capture the user's dynamic interest preferences with the codec constructed by one-dimensional CNN,and the personalized recommendation list of users is generated.Compare the contents in the user personalized recommendation list with the contents in the edge node cache list.For the contents not in the cache list,the contents in the user personalized recommendation list should be replaced as much as possible to the contents in the cache list without deviating from the upper limit of the user personalized interest preference feature vector condition.The simulation results show that the proposed strategy can increase the probability of the user's request for the cached contents of the edge node,and further improve the cache hit rate.