K-Means for Search Results Clustering Using URL and Tag Contents

Abstract

Increasing volume of web has resulted in the flooding of huge collection of web documents in search results creating difficulty for the user to browse the necessary document. Clustering is a solution to organize search results in a better way for browsing. It is a process of combining similar web documents into groups. For web page clustering, terms (features) can be extracted from different parts of a web page. Giansalvatore, Salvatore and Alessandro have extracted terms from entire web page for clustering Stanis law Osinski et al., have considered terms only from snippets. A new method is introduced in this paper which extract terms from URL, Title tag and Meta tag to produce clusters of web documents. The reason for selecting these parts of a web page is that they contain keywords which are available in a web page. Clustering algorithm used in this paper is K-means. Proposed method of clustering is compared with snippet based clustering in terms of intra-cluster distance and inter-cluster distance.