Research On Ontology-Based Automatic Annotation For Deep Web

With increasing the date of the deep web,the deep web contains the data retrievaling which has become particularly important. However,the current mainstream search engines basically only collected the Publicly Indexable Web which can be an index of information on the Internet. In fact most of the information for the traditional search engine is not visible. In order to make search engine queries for the deep web more efficient. first, We use the data annotation technology to aonnotate the date of query interface,then annotation the information Of the results page and submitted it to search engines in order to facilitate later retrieval and extraction.This annotation method uses the concept of the ontology that greatly improves consistency of theFirst of all, according to the structure of the web page to extract the information of the web page.In the extraction process is based on the spatial relationship between data information and explanation information, and if these two information in a straight line and in between these two messages there is no other information we consider that these two information match. This shows that the information is used to describe the data.After completing the extraction process,this information to mark the corresponding data.Sometimes,however, some of the results page contains the number of the data volume is very small,it is very hard if only using the information of the result page to annotate.It is necessary to use information of query interface to annotate the data in the results page. In this process, in order to ensure consistency of the data we use the information of integration interface to annotate the corresponding data.Finally,identify each type of annotation information,Annotation information can be divided into data types and text-type. this classification process allows calculation the similarity between the annotation information and the ontology that are more convenient.finally using the word by word comparing method to calculate the similarity between ontology and annotation information, and use the appropriate ontology phrase to replace the corresponding annotation information to annotate the data.