Previous abstract | Contents | Next abstract
This paper presents a named entity classification system that utilises both orthographic and contextual information. The random subspace method was employed to generate and refine attribute models. Supervised and unsupervised learning techniques used in the recombination of models to produce the final results.