Machine Learning-Based Web Application for ADHD Detection in Children

• Autores: Porras, Diego Oscar Alexander, Mejia, Gerson Antonio, Castañeda, Pedro Segundo
• Formato: artículo
• Fecha de Publicación: 2024
• Institución: Universidad Peruana de Ciencias Aplicadas
• Repositorio: UPC-Institucional
• Lenguaje: inglés
• OAI Identifier: oai:repositorioacademico.upc.edu.pe:10757/676002
• Enlace del recurso: http://hdl.handle.net/10757/676002
• Nivel de acceso: acceso embargado
• Materia: ADHD detection; Child mental health; Computing methodologies; Machine learning

Attention deficit hyperactivity disorder (ADHD) represents a medical condition characterized by the presence of inattention, hyperactivity, and impulsivity, which affects the academic development of students globally. In Peru, it affects a proportion of the pediatric population ranging from 2% to 12%, with a prevalence of 12.1% in South Lima, particularly in public schools. This research presents an online application with machine learning to improve the detection of ADHD in elementary school children. Several machine learning algorithms were reviewed and Random Forest was selected as the best-performing model with an accuracy of 96.08%. The model uses 27 selected variables, optimizing data collection and training. The child answers the questionnaire within the app and psychologists can access the app to visualize the results, aiding in the early detection of ADHD. The experiment involved 189 participants, resulting in a high accuracy of the Random Forest model. This innovative solution can have a significant impact on the early identification of ADHD, benefiting children's health and educational process.