Hybrid approach for automatic evaluation of emotion elicitation oriented to people with intellectual disabilities

  • Authors:

  • R. Martínez;K. López de Ipiña;E. Irigoyen;N. Asla

  • Affiliations:

  • Departamento Ingeniería de Sistemas y Automática, Grupo de Inteligencia Computacional;Departamento Ingeniería de Sistemas y Automática, Grupo de Inteligencia Computacional;Departamento Ingeniería de Sistemas y Automática, Grupo de Inteligencia Computacional;Departamento de Psicología Social y Metodología de las Ciencias del Comportamiento, de la Facultad de Psicología, Universidad del País Vasco/Euskal Herriko Unibertsitatea

  • Venue:
  • HAIS'10 Proceedings of the 5th international conference on Hybrid Artificial Intelligence Systems - Volume Part I
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

People with intellectual disabilities and elderly need physical and intellectual support to ensuring independent living This is one of the main issues in applying Information and Communication Technology (ICT) into Assistive Technology field In this sense the development of appropriated Intelligent Systems (ISs) offers new perspectives to this community In our project a new IS system (LAGUNTXO) which adds user affective information oriented to people with intellectual disabilities has been developed The system integrates a Human Emotion Analysis System (HEAS) which attempts to solve critical situations for this community as block stages In the development of the HEAS one of the critical issues was to create appropriated databases to train the system due to the difficulty to simulate pre-block stages in laboratory Finally a films and real sequences based emotion elicitation database was created The elicitation material was categorized with more actual features based on discrete emotions and dimensional terms (pleasant, unpleasant) Classically the evaluation is carried out by a specialist (psychologist) In this work we present a hybrid approach for Automatic Evaluation of Emotion Elicitation databases based on Machine Learning classifiers and K-means clustering The new categorization and the automatic evaluation show a high level of accuracy with respect to others methodologies presented in the literature.