Automatic document clustering using topic analysis

Abstract

Web users are demanding more out of current search engines. This can be noticed by the behaviour of users when interacting with search engines [12, 28]. Besides traditional query/results interactions, other tools are springing up on the web. An example of such tools includes web document clustering systems. The idea is for the user to interact with the system by navigating through an organised hierarchy of topics. Document clustering is ideal for unspecified search goals or for the exploration of a topic by the inexpert [21]. Document clustering is there to transform the current interactions of searching through a large amount of links into an efficient interaction where the interaction is navigation through hierarchies. This report will give an overview of the major work in this area, we will also propose our current work, progress and pitfalls which are being tackled.