Model for the Presumptive Diagnosis of Mental Disorders in Adults Applying Machine Learning

Velasquez E.D.; Barja E.R.; Wong L. · International Conference on Electrical, Computer and Energy Technologies, ICECET 2023 (2023) · DOI: 10.1109/icecet58911.2023.10389325

Resumen

Timely assessment of individuals' mental health is a global challenge that has grown over the years, often linked to exposure to various traumatic situations resulting from environmental and physical factors. In some countries, this has led to suicidal and violent behaviors, highlighting the crucial role mental health plays in society. While attempts at solutions exist, few fully harness the potential of technology. This study proposes a model for the presumptive diagnosis of mental disorders in adults using three Machine Learning (ML) algorithms: Logistic Regression, Decision Tree, and Random Forest. The model development process involves three stages: analysis, design, and validation. The first stage evaluates machine learning architectures, while the second details the design of components for the proposed ML architecture. The third stage involves simulating the ML architecture and validating it with psychologists for system effectiveness and with psychological derivation experts for indicator quality. Results from the initial experiment demonstrated that the 'accuracy' of the Random Forest algorithm scored an average of 0.93 (high level) with an average precision between 0.90 and 0.96.

Instituciones: Universidad Peruana de Ciencias Aplicadas