Detail výsledku
Measuring Web Page Similarity Based on Textual and Visual Properties
Measuring web page similarity is a very important task in the area of web mining and information retrieval. This paper introduces the method for measuring web page similarity, which considers both textual and visual properties of pages. Textual properties of a page are described by means of modified weight vector space model. General visual properties are captured via segmentation of a page, which divides a page into visual blocks, properties of which are stored into a vector of visual properties. These both vectors are then used to compute the whole web page similarity. This method will be described in detail and results of several experiments are also introduced in this paper.
Web page similarity, clustering, vector space model, vector distance, term weighting, visual blocks.
@inproceedings{BUT76500,
author="Vladimír {Bartík}",
title="Measuring Web Page Similarity Based on Textual and Visual Properties",
booktitle="The 11th International Conference on Artificial Intelligence and Soft Computing",
year="2012",
series="Lecture Notes in Artificial Intelligence, Vol. 7268",
journal="Lecture Notes in Computer Science",
number="7268",
pages="13--21",
publisher="Springer Verlag",
address="Zakopane",
isbn="978-3-642-29349-8",
issn="0302-9743",
url="https://www.fit.vut.cz/research/publication/9850/"
}