EARLY DETECTION OF MENTAL HEALTH DISORDERS USING SVM-BASED FUSION OF WEARABLE BIOSENSORS AND MOBILE HEALTH DATA
DOI: https://doi.org/10.65725/JCISE/1/1/003
JOURNAL OF COMPUTATIONAL INTELLIGENCE SCIENCE AND ENGINEERING (JCISE) Volume 1 Issue 1, July 2025

Abstract
This study proposes a machine learning framework for early detection of mental health disorders by integrating multimodal physiological data from wearable biosensors with self reported behavioral metrics collected via mobile health applications. Leveraging a Support Vector Machine (SVM) classifier, the system analyzes real-time electrodermal activity (EDA), heart rate variability (HRV), sleep patterns, and user-reported mood/lifestyle logs to identify latent biomarkers associated with depression and anxiety disorders. The SVM model, trained on historical datasets with clinically validated labels, demonstrates superior sensitivity in detecting subclinical deviations from baseline physiological states—enabling proactive intervention. Comparative evaluation against traditional diagnostic methods (e.g., clinical interviews, questionnaires) reveals a 22% improvement in early-stage disorder identification accuracy (p < 0.01) while reducing false positives through kernel-optimized feature space separation. The system’s edge-computing compatibility addresses privacy concerns by enabling on-device analysis without continuous cloud dependency.

Authors: Dr.C. Kalpana, Dr.Anita Manoj Devare

Keywords: Self-Reported Mood; Physiological Data; Integrated System; Traditional Methods; Health Disorders; Vector Machine