Pengembangan Sistem Prediksi Risiko Gangguan Mental Remaja Menggunakan Support Vector Machine (SVM)

Abstract

Adolescent mental health has become an increasingly critical issue due to the rising prevalence 
of emotional and behavioral disorders among young individuals. Social pressure, academic demands, and 
psychological changes often trigger stress, anxiety, and even depression, which affect learning activities 
and social interactions. This study aims to develop a web-based system to detect mental disorder risk in 
adolescents using a machine learning approach with the Support Vector Machine (SVM) algorithm. Three 
open datasets from the Kaggle platform—Big Five Personality Test Dataset, Symptom2Disease Dataset, and Mental Health in Tech Survey Dataset—were utilized to integrate personality traits, physical 
conditions, and mental health indicators. The data underwent preprocessing involving duplicate removal, 
missing value imputation, standardization, and categorical-to-numerical transformation before being split 
into 70% training and 30% testing sets. The system was developed using the Agile Scrum methodology in 
an iterative and adaptive manner based on user feedback. The experimental results show that the SVM 
model with an RBF kernel achieved 91.3% accuracy, 89.7% precision, and 91.9% F1-score. The resulting 
system, can classify mental disorder risk levels and provide prevention recommendations according to the 
assessment results. With an interactive interface, this system is expected to assist adolescents in recognizing 
their mental conditions early, increase awareness of psychological well-being, and serve as a technology
based educational tool for mental health prevention.