Machine learning-based detection techniques for post-traumatic stress syndrome recognition

Post-Traumatic Stress Disorder (PTSD) is regarded as a medical condition that affects a large number of individuals worldwide from a variety of professions. However, PTSD diagnosis and acquiring precise diagnosis methods have become more difficult tasks. In this research, a novel Machine Learning (ML) system, the Hybrid Glow-Worm Optimized Multi-Level Spectral Support Vector Machine (HGO-MLSSVM), is proposed for PTSD recognition. The primary objective of the proposed method is to extract and classify the voice frame. The audio dataset was employed to assess and categorize the speech signal-derived PTSD. Following that, pre-processing, feature extraction, and classification are processed to detect PTSD patients through the HGO-MLSSVM technique. The proposed method produces very promising results, obtaining ROC (0.98), specificity (98.02%), sensitivity (97.5%), and accuracy (98.50%) when compared to traditional approaches. Furthermore, this study shows the significant effectiveness of the HGO-MLSSVM technique compared to current methodologies and a discourse on prospective directions for forthcoming PTSD recognition.