Rule induction algorithms for classification of psychotic disorders involving social vulnerability features

Author(s)

Carvalho, Isabelle ; Garcia Rosa, Joao Luis ; dos Santos, Ketlin Fabri ; Alves, Domingos ; QuintelaVarajao, JE ; CruzCunha, MM ; Martinho, R ; Rijo, R ; Domingos, D ; Peres, E

• Document type: Journal article
• Source: PROCEEDINGS OF THE XI LATIN AND AMERICAN ALGORITHMS, GRAPHS AND OPTIMIZATION SYMPOSIUM; v. 138, p. 7-pg., 2018-01-01.

Abstract

This article presents rule induction algorithms for identify and evaluate variables related to social vulnerability risk factors of the individual that influence the diagnosis of psychotic disorders. 213 variables were analyzed: sociodemographic, psychic experiences in the community, traumas in childhood, bullying, and use of marijuana, tobacco, and alcohol. Four subsets were constructed and with them, we reach an attribute reduction of 80%. In development, an accuracy of approximately 80% was maintained. About the number of rules there was a reduction of 90% of possibilities generated, simplifying the representation of the model. Thus, the results proved promising for the use of this approach as a basis in the construction of decision support systems in the area and in the deepening of studies for the incorporation of new variables in psychiatric diagnosis, increasing the reliability and validity of the diagnostic process. (C) 2018 The Authors. Published by Elsevier Ltd. (AU)

• FAPESP's process: 16/02555-8 - Development of algorithms and computational techniques for application in brain-computer interfaces
• Grantee: João Luís Garcia Rosa
• Support Opportunities: Regular Research Grants