Enhancing web page classification through image-block importance analysis

Elisabetta Fersini

doi:10.1016/j.ipm.2007.11.003

2008-07-01

PRE

Elisabetta Fersini *

Enza Messina

Francesco Archetti

DOI：10.1016/j.ipm.2007.11.003

原文链接

原文求助

摘要

En 中文

We present a term weighting approach for improving web page classification, based on the assumption that the images of a web page are those elements which mainly attract the attention of the user. This assumption implies that the text contained in the visual block in which an image is located, called image-block, should contain significant information about the page contents. In this paper we propose a new metric, called the Inverse Term Importance Metric, aimed at assigning higher weights to important terms contained into important image-blocks identified by performing a visual layout analysis. We propose different methods to estimate the visual image-blocks importance, to smooth the term weight according to the importance of the blocks in which the term is located. The traditional TFxIDF model is modified accordingly and used in the classification task. The effectiveness of this new metric and the proposed block evaluation methods have been validated using different classification algorithms. (C) 2007 Elsevier Ltd. All rights reserved.

Keyword:

term weighting

vector space model

visual layout analysis

document classification