Automatic Web Page Classification
Jiří Materna · RASLAN · 2008
Aim of this paper is to describe a method of automatic web page classification to semantic domains and its evaluation. The classifica- tion method exploits machine learning algorithms and several morpho- logical as well as semantical text processing tools. In contrast to general text document classification, in the web document classification there are often problems with short web pages. In this paper we proposed two ap- proaches to eliminate the lack of information. In the first one we consider a wider context of a web page. That means we analyze web pages refer- enced from the investigated page. The second approach is based on so- phisticated term clustering by their similar grammatical context. This is done using statistic corpora tool the Sketch Engine.