To solve the problem of seriousinformation overload caused by the explosive growth of Internet information

atime-weightednetwork structure recommendation algorithm is proposed by studying thetraditional recommendation algorithm. This algorithm creates a "User-object-tag"tripartite graph and uses the mass diffusion algorithm to calculate theresource vectors obtained of the project. Then

according to the timeattenuation problem

we design the time-weighted function based on usersocial tags. Finally

a new network structure recommendation algorithm isproposed by combining the time-weighted function with the tripartite graph recommendationalgorithm. the experimental results demonstrate that the recommendationalgorithm can solve the problem of dynamic migration about user interests andsignificantly improve accuracy of recommendation effectively.