Research On Deep Web Sources Classification

According to the Internet depth can be divided into Deep Web and Surface W eb two class. Deep Web contains information compared to the Surface Web inform ation more useful. However, the Deep Web information has independent forms are distributed across a network, it is a dynamic, constantly changing, therefore, this bri ngs difficulties to get the Deep Web information. In order to effectively provide th e information, must carry on the data integration of Deep Web.The Deep Web query interface determining. Before the web form feature extrac tion joined the heuristic rules. The use of classifier integration thought, using Adab oost algorithm of multiple simple Bias classifier are accumulated to form a strong classifier, weaken the request attribute independence assumption of the naive Bias c lassification algorithm, the whole body to improve the classifier performance by usi ng the difference between multiple classifier. The experimental results proved the g ood judgment effect. The Deep Web data source classification. In the classification of the data source, using the KNN classification algorithm to. Because the KNN c1 assification algorithm K value selected is too large or too small will affect the clas sification results, so the algorithm is optimized and verified by experiments to obtai n good results.