Database Technology on the Web: Query Interface Determining Algorithm for Deep Web Based on HTML Features and Hierarchical Clustering

摘要: According to the features of Hypertext Markup Language, interactive elements exist in terminal Document Object Model tree and they are close each other local area, we proposed a method find web query interface which combines models rules. In this method, after establishing model locate parts interfaces by interaction density cluster groups their similarity structure hierarchically. Then some nonquery interfaces filtered out help content-filter composed This avoids excessive dependence on tag “form” presents better performance than traditional methods property accuracy generality. And accuracy of experiment results on common dataset TEL-8 self-organized reached respectively 90.1% 92%.