Melampus: um modelo deep learning para triagem psicológica infantil

Abstract

While mental health related issues usually began in childhood or adolescence, only a small portion of this population receives proper diagnosis and treatment. In part, this scenario is caused due the lack of specialized tools for mental disorders screening, mainly those which reduce cost and time needed. Recently, authors have been analyzing how machine learning could help to build new psychological assessment tools. However, only few researches proposed building specialized tools for groups composed mostly of children. This work aims to propose a model which combines clinical tests and deep learning for supporting child psychological screening. Results suggested that deep learning tools may fit the intended scenario, once the classification models tested performed well even with a small sample size.