A new approach for the discovery of web communities

Abstract EN

During the past few years the World Wide Web size has grown exponentially.

Search engines are estimated to index billions of web pages.

Users of the World Wide Web usually are faced with numerous and irrelevant search results returned from the search engines.

Community detection over the World Wide Web is one of the important activities in the web.

Web community detection is typically concerned with discovering related web pages focused around a certain topic.

Traditional methods for web community detection include searching the link-structure of the web, using graph theory or using maximum flow algorithm.

This paper suggests the usage of association rule mining in the detection of web communities.

The main idea is to identify the web communities by transforming a sub-graph around a certain search topic of a web graph into a database of transactions.

We evaluated our approach by running several experiments.

This is achieved by examining the performance and functionality of the proposed algorithm.

Comparison with another known algorithm will be presented.