Personalized information recommendation based on synonymy tag optimization

Synonymy tags are very common in social tagging system, which also influence the performance of recommendation algorithm based on tags. In order to obtain a better performance, this paper uses WordNet to determine the meaning of target tags, and consturcts synonymy tag set for collecting synonymy tags under a defined classification, which is a result of application Clique Percolation Method on user saved resources. Then, the set is applied to resource model for extension, and weight coefficient method is adopted while the synonymy tags are absorbed in. On basis of this, the recommendation algorithm is put forward, and dataset from hetrec is taken for experiment. A new evaluation method named quality value of recommended resource is created in the paper, and the results show that the performance of optimization algorithm is better than the baseline one, but with the increasing number of users and resources, the percentage of improvement is become narrowed.