SciELO - Scientific Electronic Library Online

vol.10 número2Un método para la generación de rankings en la selección de equipos de trabajo en ambiente competitivo basado en algoritmos genéticos índice de autoresíndice de materiabúsqueda de artículos
Home Pagelista alfabética de revistas  

Servicios Personalizados




  • No hay articulos citadosCitado por SciELO

Links relacionados

  • No hay articulos similaresSimilares en SciELO


Revista Cubana de Ciencias Informáticas

versión On-line ISSN 2227-1899


RODRIGUEZ VAZQUEZ, Solangel  y  MARTINEZ BORGES, Andy Vidal. Evaluation of alternatives for cervical cell classification using only nucleus’ features. Rev cuba cienc informat [online]. 2016, vol.10, n.2, pp.211-222. ISSN 2227-1899.

This paper presents a comparative study of the performance of three of the most widely used classification techniques for diagnosis in Pap tests based solely on features extracted from the nucleus’ region. Among the selected techniques are the nearest neighbor classifier (kNN), artificial neural networks (ANN-RBF Network) and support vector machines (SVM). The comparative study is performed in order to determine the technique with greater ability to correctly classify the patterns that identify changes in cervical cells from feature matrices from images of Papanicolaou smears. In this study the results obtained from the iterations in each classifier using the same data sets were used in order to determine, which method offers the best solution to the problem of cervical cells classification. An experimental study was conducted considering domains with equal numbers of classes, attributes, and training examples, as well as an equal proportion of cases belonging to each class. The statistical comparison was made based in the results obtained in terms of the indexes of effectiveness known as F-measure, H-mean, negative predictivity and area under the ROC curve (AUC).

Palabras clave : cervical cells; Papanicolaou test; classification; kNN; neural networks; support vector machines.

        · resumen en Español     · texto en Español     · Español ( pdf )


Creative Commons License Todo el contenido de esta revista, excepto dónde está identificado, está bajo una Licencia Creative Commons