Research On Quality Estimation Model Of Deep Web Data Sources And Application

With the rapid development of Internet technology,the Web has been rapidly deepened by myriad searchable databases online.A large amount of dynamic information from the databases behind query interfaces can not be retrieved because of the restrictions of current search engine technology.We call such information as Deep Web.Deep Web information retrieval is still a fresh field of study and has been paid more and more attention.In attempt to meet users' need for Deep Web information,the research on Deep Web information integration in large-scale has born.It including data sources finding,data sources classification,data sources selection,results combination.In this paper,we present a quality estimation model of Deep Web data sources.Based on the model we do applied research in data sources selection,and bring forward a related algorithm.The main research contents including:(1) Introduce the information integration framework of Deep Web,do research on Deep Web structure and distribution of Chinese resources.(2) Aanalysis of the three aspects characteristics of Deep Web,and extracte the characteristics of impact the quality of the data sources.(3) Based on the factors affected the quality,using machine learning method and fuzzy comprehensive evaluation method to establish data sources quality estimation model respectively.(4) Application of the quality estimation model above,combining with inquiries relation and inquiries accuracy,we do applied research on Deep Web data sources selection.Finally,we design experimente to verify the methods and techniques proposed in this papaer.The experimental results verify the extracted characteristics are reasonable and effective.And we compared the advantages and disadvantages of the estimation model, using machine learning method and fuzzy comprehensive evaluation method to establish.