PTSD Detection using BERT Models
Developed a system to detect PTSD indicators in clinical text using BERT and its variants. The project involved comprehensive data preprocessing and fine-tuning models to achieve high validation accuracies (93-95%) across multiple BERT configurations.
