Early detection of mental health disorders using machine learning models using behavioral and voice data analysis

Area of Science:

• Artificial Intelligence
• Computational Psychiatry
• Machine Learning in Healthcare

Background:

• Mental illness is a global health issue requiring early diagnosis for effective treatment.
• Manual diagnosis is time-consuming and prone to delays, increasing risks like suicide.
• AI offers potential for accurate and timely mental health disorder detection.

Purpose of the Study:

• To develop a novel AI framework, NeuroVibeNet, for early mental illness detection.
• To leverage a multi-modal approach integrating speech and behavioral data.
• To achieve high accuracy and reliability in distinguishing normal from pathological conditions.

Main Methods:

• A multi-modal framework preprocessing speech and behavioral datasets.
• Combining Improved Random Forest (IRF) and Light Gradient-Boosting Machine (LightGBM) for behavioral data.
• Integrating Hybrid Support Vector Machine (SVM) and K-Nearest Neighbors (KNN) for voice data, with a weighted voting mechanism.

Main Results:

• The NeuroVibeNet framework demonstrated robust performance.
• Achieved a competitive accuracy of 99.06% in classifying normal and pathological mental health states.
• Validated the effectiveness of multi-modal data integration for mental illness detection.

Conclusions:

• The proposed AI framework enables reliable and early detection of mental illness disorders.
• Multi-modal data integration significantly enhances diagnostic accuracy.
• NeuroVibeNet shows promise for improving mental healthcare outcomes through timely intervention.