Semantic Web Prefetching Using Semantic Relatedness between Web pages

Semantic web prefetching makes use of the semantics of the keywords provided by the user for making predictions. These keywords can be descriptive texts like anchor texts, texts surrounding anchor texts, titles or simple text of the content. For finding the relation between two webpages, semantic information can be used. Semantic information is embedded within the web pages during their designing for the purpose of reflecting the relationship between the web pages. The client can fetch this information from the server. However, this technique involves load on web designers for adding external tags and on server for providing this information along with the desired page, which is not desirable. We are trying to find the semantic relation between web pages using the keywords provided by the user and the anchor texts of the hyperlinks on the present web page. Various semantic relations are: sequential (seq), similar (sim), cause-effective (ce), implication (imp), subtype (st), instance (ins), referential (ref). Their priorities are in the order: seq>sim>ce>imp>st>ins>ref. This paper presents algorithms for finding sequential and similar relations between web pages in order to improve the performance of web prefetching with no overhead on designers and load on server. These algorithms are to be implemented on the client side. These algorithms do not involve any overhead of web designers and extra load on server.