Abstract
Depression is a widespread global disease of increasing global concern. Early recognition of signs of depression is crucial to evaluating and treating or preventing mental illness. With advances in machine learning, it has become possible to develop intelligent systems capable of recognizing depression and its signs in speech by analyzing speech and processing audio signals. This study presents an AI model for detecting and predicting mental illnesses through speech analysis of medical datasets related to depression. We used Distress Analysis Interview Corpus-Wizard of Oz (DAIC-WOZ) database where 60% of the data was reserved for training, while 20% was allocated for testing the model and another 20% for model validation. The model includes a convolutional neural network (CNN) to detect and predict mental illnesses. The proposed CNN model achieved an accuracy of 82% in the training and testing phases. Ultimately, the results are useful for classifying depression during English speaking and will be useful to psychiatrists and psychologists in contributing to early detection of depression at an affordable cost.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call