Tree-based Classification of EEG Spectra

Abstract

A novel approach to EEG spectra classification aimed at somnolence detection is presented. Classification forest based on data of a single individual is combined with a "global" classification model derived from data of a group of subjects. The combined model attains a higher mean accuracy than both of its components.

Categories and tags

  • Konferenční příspěvek (mezinárodní konf.)