Neural Networks for Web Content Filtering

  • Authors:
  • Pui Y. Lee;Siu C. Hui;Alvis Cheuk M. Fong

  • Affiliations:
  • -;-;-

  • Venue:
  • IEEE Intelligent Systems
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

The proliferation of objectionable material on the Internet has created an urgentneed for countermeasures to protect unsuspecting children and others from such material'sharmful effects. However, current Web content-filtering techniques and commerciallyavailable Web-filtering systems have serious shortcomings. Machine intelligence cancompensate for these shortcomings. For example, the Intelligent Classification Engine usesneural networks' learning capabilities to provide fast, accurate differentiation betweenpornographic and nonpornographic Web pages. The engine works with both Kohonen'sSelf-Organizing Maps and Fuzzy Adaptive Resonance Theory networks. Both networks performsignificantly better than nonintelligent techniques; KSOM has greater classificationaccuracy, but training it takes longer.