Optimization-based methods for classification and regression problems with imprecise data

Tesis doctoral de José Feliciano Gordillo Santofimia

In this thesis, we develop several tools to solve classification and regression problems where the elements of the database are sets affected by some kind of uncertainty several particular cases are studied via this methodology, interval data. Perturbed data and bags of instances. the classifier and regressor is defined by following the strategy successfully used in support vector machines of maximizing the margin in each case, an optimization problem is formulated which must be solved to obtain an optimal classifier or regressor. these techniques are shown to be useful for imputation based on intervals and for a semi-obnoxious location problem (in the framework of locational analysis).

 

Datos académicos de la tesis doctoral «Optimization-based methods for classification and regression problems with imprecise data«

  • Título de la tesis:  Optimization-based methods for classification and regression problems with imprecise data
  • Autor:  José Feliciano Gordillo Santofimia
  • Universidad:  Sevilla
  • Fecha de lectura de la tesis:  28/03/2008

 

Dirección y tribunal

  • Director de la tesis
    • Emilio Carrizosa Priego
  • Tribunal
    • Presidente del tribunal: eduardo Conde sanchez
    • belen Martin barragan (vocal)
    • bernard Menderick (vocal)
    • edgard Nyssen (vocal)

 

Deja un comentario

Tu dirección de correo electrónico no será publicada. Los campos obligatorios están marcados con *

Scroll al inicio