Abstract:

In order to solve the inability of personalized recommendation system on the semantic information processing, the semantic recommendation system model was built, which can descript the semantic sence in users interest information. A kind of recommendation algorithm was also proposed to calculate the relationship between users and resource. Description logic was used to apply the model in a special domain, two rules was introduced to achieve the ability of interest transmission between different level of semantic concepts both in user profiles and resource profiles. The experiment shows that semantic recommendation system model could produce more related results for particular user than classic methods, it can discovery more new interest which is implicated in the interest concepts.

Key words: personalized recommendation, description logic, system modeling