A dynamic information retrieval system for the Web

Abstract

In conventional information retrieval (IR), documents in a collection are indexed before the retrieval process, and the document collection is generally static, organised and homogeneous. But, information on the Web is vast, dynamic, unorganised and heterogeneous. Current search engines and IR systems on the Web are based on the conventional indexing approach and have limitations including the need for frequent update of the index. We propose a non-indexing approach for information retrieval on the Web and show that the IR system based on this approach performs better than popular search engines.