Automatic Web Content Extraction for Generating Tag Clouds from Thai Web Sites

Abstract

This paper proposes a novel Web content extraction approach based on heuristic rules and the XPath utility in XML. The main objective is to address the problem of Web visualization by generating tag clouds from Thai Web sites in order to provide an overview of the key words in the Web pages. This paper also proposes a detailed method to assess the Web content extraction technique on a single Web page by using the length of the extracted content. There are three main steps in the proposed technique: Web page elements and features extraction, Block detection, and Content extraction selection. The empirical results have shown this technique produces high accuracies.