Knowledge Discovery from Semistructured Texts

  • Authors:
  • Hiroshi Sakamoto;Hiroki Arimura;Setsuo Arikawa

  • Affiliations:
  • -;-;-

  • Venue:
  • Progress in Discovery Science, Final Report of the Japanese Discovery Science Project
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper surveys our recent results on the knowledge discovery from semistructured texts, which contain heterogeneous structures represented by labeled trees. The aim of our study is to extract useful information from documents on the Web. First, we present the theoretical results on learning rewriting rules between labeled trees. Second, we apply our method to the learning HTML trees in the framework of the wrapper induction. We also examine our algorithms for real world HTML documents and present the results.