Knowledgebase Harvesting for User-Adaptive Systems Through Focused Crawling and Semantic Web
Bujar Raufi, Florije Ismaili, Jaumin Ajdari, Xhemal Zenuni · 2016
The expansion and ever evolving web makes it difficult to find relevant information that best fits user's intentions. This paper introduces development of hybrid approach that addresses the issue of collecting large knowledgebase by fusing the thematic or focused crawling methodologies, with adaptive and semantic web concepts. Focused crawling ensures the goal directed search of data on the web, adaptive web environments establish proper content and link adaptation whilst semantic web inserts meaning to web documents from the sense of content and metadata. The thematic crawling process retrieved approximately 11,429 documents from 11,286 visited locations resulting in 9,807 database entries out of which 81 entries are classified as top ranked distributed in seven categories. On the next phase, a reasoning process was performed against the semantic ontology which comprised the top ranked documents as class individuals. Results indicated that retrieved relevant documents, together with assertions against class individuals from the ontology highly reflect the user browsing activities and intentions.