Measuring the error of linear separators on linearly inseparable data

• Boris Aronov ^[5] ; Delia Garijo ^[1] ; Yurai Núñez Rodríguez ^[2] ; David Rappaport ^[2] ; Carlos Seara ^[3] ; Jorge Urrutia ^[4]
  1. [1] Universidad de Sevilla

     Universidad de Sevilla

     Sevilla, España

     
  2. [2] Queen's University

     Queen's University

     Canadá

     
  3. [3] Universitat Politècnica de Catalunya

     Universitat Politècnica de Catalunya

     Barcelona, España

     
  4. [4] Universidad Nacional Autónoma de México

     Universidad Nacional Autónoma de México

     México

     
  5. [5] Polytechnic University

  Mostrar afiliaciones +
• Localización: XIII Encuentros de Geometría Computacional: Zaragoza, del 29 de junio al 1 de julio de 2009 / Alfredo García Olaverri (ed. lit.) , Javier Tejel Altarriba (ed. lit.) , 2009, ISBN 978-84-92774-11-1, págs. 251-258
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • Given a set of red points R and blue points B we seek to determine a linear separator of the sets, although the sets may be linearly inseparable. Thus we obtain a partition into two parts, where each part may contain some misclassified points. We determine the error of the partition based on the amount of work needed to move the misclassified points. We consider several different measures of work and provide algorithms to find linear separators that minimize the error under these different measures.