SciELO - Scientific Electronic Library Online

vol.8 issue1The hard objectivity of the essay exams in academic achievement assessmentGraduates' perceptions about the acquired clinical competences in undergraduate study of medicine in National University of the Northeast, UNNE, Argentina author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand




Related links

  • On index processCited by Google
  • Have no similar articlesSimilars in SciELO
  • On index processSimilars in Google


Educación Médica

Print version ISSN 1575-1813


BORRACCI, Raúl A.  and  ARRIBALZAGA, Eduardo B.. Cluster analysis and artificial neural networks for residency candidates classification and selection. Educ. méd. [online]. 2005, vol.8, n.1, pp.22-30. ISSN 1575-1813.

Introduction: Multiple linear regression models, computer-assisted selection and recently, artificial neural networks have been used at educational programs to produce preliminary rank lists of residency applicants. The aim of this study was to evaluate and redesign a system to rank applicants for a university residency program using multivariate analysis and neural networks models. Methods: The design was a retrospective-transversal study, performed in University Hospital. A random sample of 213 residency applicants to a medical university program was evaluated with regard to medical school grades, examinations, autobiography, internship and interview scores. Hierarchical cluster análisis and artificial neural networks for applicants’ classification and ranking were developed using standardized scores of all 5 variables. Results: Cluster analysis classified applicants in 12 clusters depending on average standardized values of variables. This analysis was used to construct a descriptive classification of groups and a final ranking list according to applicant’s relative position over or under average scores. Multi-layer perceptron network was able to imitate the cluster solution with a mean sensitivity and specificity level of 94.1% and 99.1% respectively. Conclusions: A hierarchical cluster analysis was used to classify a sample of residency applicants in a ranking list, based on candidate’s relative position over or under the mean standardized scores of individual variables. Additionally, a MLP network was trained to imitate cluster results with a sufficient level of accuracy to be considered as an optional computer-assisted method to cluster analysis of massive data. This cluster solution constitutes an alternative approach for residency candidates’ selection.

Keywords : residency; cluster analysis; artificial neural network.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )


Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License