Author(s)

G. M. Caputo & N. F. F. Ebecken

Abstract

This work presents a study of web communities, discovering links and associations between business web pages. Those links are drawn using graphs where the nodes are mapped, entities representing business subjects, such as products, services, enterprises name, partnerships, and some characteristics evolving those entities. The edges are the relationships between those mapped entities. In comparing the resulting graphs it could be expected to find differentiations between competitors and their strategies. To achieve this objective, communities’ discovery techniques and visual analysis are applied and tested using a set of graph representations and different web crawlers’ depths. On further analysis, the obtained bipartite graph aims to discover market trends and help competitive intelligence analysts. Keywords: link mining, web mining, group detection.

Keywords

link mining, web mining, group detection.