A Personalization Recommendation Method Based on Deep Web Data Query

Deep Web is becoming a hot research topic in the area of database. Most of the existing researches mainly focus on Deep Web data integration technology. Deep Web data integration can partly satisfy people's needs of Deep Web information search, but it cannot learn users' interest, and people search the same content online repeatedly would cause much unnecessary waste. According to this kind of demand, this paper introduced personalization recommendation to the Deep Web data query, proposed a user interest model based on fine-grained management of structured data and a similarity matching algorithm based on attribute eigenvector in allusion to personalization recommendation. Secondly, As for Deep Web information crawl, a crawl technology based on the tree structure is presented, with the traversal method of tree to solve the information crawl problems in the personalization service distributed in various web databases. Finally, developed a prototype recommendation system based on recruitment information, verified the efficiency and effectiveness of the personalization recommendation and the coverage and cost of Deep Web crawl through the experiment.