Back to Search Start Over

Using ant-behavior-based simulation model AntWeb to improve website organization

Authors :
Weigang Li
Wesley Martins Teles
Marcos Vinícius Pinheiro Dib
Vlaudemir Morais de Andrade
Judas Tadeu Cariolano
Alba Cristina Magalhaes Alves de Melo
Source :
Data Mining and Knowledge Discovery: Theory, Tools, and Technology
Publication Year :
2002
Publisher :
SPIE, 2002.

Abstract

Some web usage mining algorithms showed the potential application to find the difference among the organizations expected by visitors to the website. However, there are still no efficient method and criterion for a web administrator to measure the performance of the modification. In this paper, we developed an AntWeb, a model inspired by ants' behavior to simulate the sequence of visiting the website, in order to measure the efficient of the web structure. We implemented a web usage mining algorithm using backtrack to the intranet website of the Politec Informatic Ltd., Brazil. We defined throughput (the number of visitors to reach their target pages per time unit relates to the total number of visitors) as an index to measure the website's performance. We also used the link in a web page to represent the effect of visitors' pheromone trails. For every modification in the website organization, for example, putting a link from the expected location to the target object, the simulation reported the value of throughput as a quick answer about this modification. The experiment showed the stability of our simulation model, and a positive modification to the intranet website of the Politec.

Details

ISSN :
0277786X
Database :
OpenAIRE
Journal :
SPIE Proceedings
Accession number :
edsair.doi...........57dd9c7873066a143183b8075f4b7430