Classification and Characterization of Children and Adolescents with Depressive Symptomatology using Machine Learning
| Autor: | Thiago Lima, Maycoln Leôni Martins Teodoro, Felipe Salgado, Cristiane Neri Nobre, Kelly Malaquias, Renata Cristina Santana |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
020205 medical informatics
Artificial neural network business.industry Decision tree 02 engineering and technology Machine learning computer.software_genre medicine.disease Support vector machine 03 medical and health sciences 0302 clinical medicine Multilayer perceptron 0202 electrical engineering electronic engineering information engineering medicine Major depressive disorder Prognostics Artificial intelligence Precision and recall Psychology business computer 030217 neurology & neurosurgery Depression (differential diagnoses) |
| Zdroj: | SMC |
| DOI: | 10.1109/smc.2019.8914067 |
| Popis: | According to the World Health organization (WHO), there are currently in the world more than 300 millions of people living with depression symptoms. Depression is a disorder that results from a complex interaction of biological, psychological, and social factors, and it is known for having difficulties in both diagnostics and prognostics. Machine Learning techniques are increasingly and often used to classify or characterize different profiles of diseases. This paper presents a study about major depressive disorder among Brazilian’s children and adolescents by using decision trees classifiers, Support Vector Machine (SVM) and Neural Network. A discussion about the identified attributes is presented, including, for instance, the great relation of suicidal thoughts with elevated symptomatology. Beyond that, the value of the evaluation method F-Measure, the weighted harmonic mean of precision and recall, was above 88% for both classes, high and low symptomatology for depression; which reached values above 95% when used Multilayer Perceptron and SMO algorithms, they are based in Neural Networks and Support Vector Machine, respectively. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|