The Study On Deep Web Interface Integration And Search Strategy

With the rapid development of computer networks and information technology, the rapid increase of the information on the web has become one of the most important social information resources. Nowadays, more and more people are depending on search engine to obtain their information. However, there are many online databases in the Web, called Deep Web. These databases'information is real-time built through inquiring searching, but it is not visual for main search engines. The information in Deep Web is much bigger in scale and higher in quality. Therefore, the establishment of Deep Web data integration system has become a hot issue of investigation in the domain of database and information search.We carry out research in the respect of Deep Web query interface integration and the search strategy of unified query interface. They are the important content of Deep Web data integration study.In the query interface integration, this thesis first categorizes the Deep Web interfaces through analyzing the Deep Web query interface pages'structure and their manifestation. Then we introduce the concept of interface element and give the query interface a formal description. Based on these, this dissertation proposes a method to integrate Deep Web interfaces through knowledge study and probing query. In our approach, we first select the required attributes, and then match these attributes in each interface element, after that integrate each interface element which matches the same attribute to generate the unified query interface. This method includes interface template-based matching, domain knowledge-based matching and probing queries-based matching, etc. Experimental results show that the proposed method has higher matching accuracy and lower dependency on preparatory work.For the integrated unified query interface, this dissertation improves its search strategy. First, for the different types of the integrated Deep Web query interfaces, we propose three mapping ways and the second query method to expand a unified query interface and improve query accuracy. Then we clarify a method to improve query efficiency by establishing the local index database. Analysis shows that the methods described in this dissertation have high query accuracy and time efficiency.