Speech-Based Depression Detection System Optimized Using Particle Swarm Optimization

Abstract

Depression is a prevalent mental disorder, and accurate diagnosis is essential for effective intervention. Traditional diagnostic methods rely on subjective professional judgment, which can introduce bias. To address this challenge, we propose an optimized model that hybridizes Particle Swarm Optimization (PSO), and deep-learning for depression detection and assessment. Our approach leverages Mel-frequency cepstral coefficients (MFCC), Bidirectional Long Short-Term Memory (BiLSTM), and PSO on the audio data. Experiments on the Distress Analysis Interview Corpus Wizard-of-Oz (DAIC-WOZ) dataset demonstrate significant performance improvements. Our model achieves a Root Mean Square Error (RMSE) of 3.82 and a Mean Absolute Error (MAE) of 3.24, outperforming existing state-of-the-art depression detection methods.