Hybrid dimensionality reduction approach for web page classification

Abstract

Today there is huge amount of data available on World Wide Web. One way to manage data is web page classification. One of the issues of web page classification considered in this paper is high dimensionality. Dimensionality refers to number of terms in a web page. High dimensionality of web pages causes problem while classifying them. The main objective of reducing dimensionality of web pages is to improve the performance of the classifier. This paper describes hybrid approach of dimensionality reduction for web page classification using a rough set and information gain method. Feature selection and dimensionality reduction methods are used to reduce the dimensionality of web pages. Information gain method is used as feature selection method. Rough set based Quick Reduct algorithm is used for dimensionality reduction. Web pages are classified using naïve Bayesian method. Significant results are obtained and tested for proposed architecture.

Automatic Web Page Categorization Using Machine Learning and Educational-Based Corpus

“…Sarode & Gadge [13] conducted a research that described an approach a hybrid approach for dimensionality reduction in web page classification using a rough set of naï ve Bayesian method. Dimensionality reduction is important since web pages tend to have a great number of terms and this may cause problems such as processing time.…”