A semantic-based framework for summarization and page segmentation in web mining

This chapter addresses two crucial issues that arise when one applies Web-mining techniques for extracting relevant information. The first one is the acquisition of useful knowledge from textual data; the second issue stems from the fact that a web page often proposes a considerable amount of ‘noise’ with respect to the sections that are truly informative for the user's purposes. The novelty contribution of this work lies in a framework that can tackle both these tasks at the same time, supporting text summarization and page segmentation. The approach achieves this goal by exploiting semantic networks to map natural language into an abstract representation, which eventually supports the identification of the topics addressed in a text source. A heuristic algorithm uses the abstract representation to highlight the relevant segments of text in the original document. The verification of the approach effectiveness involved a publicly available benchmark, the DUC 2002 dataset, and satisfactory results confirmed the method effectiveness.