Dimensionality reduction in a database related with viticulture crops using wrapper techniques

  1. Fernandez-Martinez, R. 1
  2. Fernandez-Ceniceros, J. 1
  3. Sanz-Garcia, A. 1
  4. Lostado-Lorza, R. 1
  5. Martinez-De-Pison-Ascacibar, F.J. 1
  1. 1 Universidad de La Rioja
    info

    Universidad de La Rioja

    Logroño, España

    ROR https://ror.org/0553yr311

Revista:
International Journal of Mathematical Models and Methods in Applied Sciences

ISSN: 1998-0140

Año de publicación: 2011

Volumen: 5

Número: 5

Páginas: 866-873

Tipo: Artículo

Otras publicaciones en: International Journal of Mathematical Models and Methods in Applied Sciences

Resumen

Recent advance in environmental monitoring technologies allows that every day, major amount of agricultural productions have a support to be controlled better. Coverall, thanks to manufacturing advances of new sensors, which allow realizing acquisition of physical variables with almost no limitation. This entails the existence of great amount of stored data, distributed in different variables that make really complicated work with them. In these circumstances, the problem arises at the time of building models when it works with a large number of variables. In order to solve it, feature selection methods are used to reduce this large number, improving building, training and validation models processes based on machine learning techniques. The methods used due to their satisfactory results, in the practical case of several viticulture crops, have been wrappers methods.