An Ontology-Based Topical Crawling Algorithm for Accessing Deep Web Content

Due to the large volume of the Web information and relatively high speed of information update, the coverage and quality of the retrieved pages by modern search engines is comparatively small. Given the volume of the Web and its frequency of content change, the coverage and quality of pages retrieved by modern search engines is relatively small since they crawl only hypertext links ignoring the search forms which are the entry points for accessing deep web content where two-thirds of information is resides. In this paper an algorithm has been designed to enable topical crawlers to access hidden web content by using domain based ontology to determine the forms' relevance to the domain. In this work scientific research publications domain has been considered. Experimental results show that proposed approach is better as compared to keyword based crawlers in terms of both relevancy and completeness.