Automated Depressive Disorder Classification using Optimized CNN-LSTM

Due to a lack of early treatment options and resources for depression detection, millions of people suffer from mental illness. It is the primary cause of sleep disorders, bipolar illness, anxiety disorders, and occasionally, self-harm and suicide. Therefore, identifying those who are experiencing mental health illnesses and getting them treated as soon as feasible is an extremely difficult endeavour. An improved deep learning Architecture (ODLA) intended for the identification and categorization of major depressive disorders is presented in this paper. The Optimized Convolutional Neural Network—Long Short-Term Memory (O-CNN-LSTM) is the name of this suggested model. The suggested methodology is applied to merge textual and auditory aspects of patient replies. The DAIC-WoZ database is used to verify the behavioural characteristics of patients with depression. The classifier is used to train the features in the proposed model. With Fire Hawks Optimization (FHO), the hyper parameters of the classifier are chosen. The suggested approach is put into practice in MATLAB, and results are assessed. It is then contrasted with traditional methods.