Abstract

A study on stress detection utilizing voice signals and two machine learning methods, Support Vector Machines (SVM) and K-Nearest Neighbours (KNN), is presented in this work. Speech-based stress detection has crucial implications in mental health support and stress management. As inputs for categorization, the system uses acoustic data collected from speech recordings. For evaluation, a dataset of voice recordings from individuals under various stress circumstances is used. The experimental results suggest that the SVM and KNN algorithms are successful at recognizing stress in voice signals. This study contributes to the field of stress detection by highlighting the potential of employing speech as a stress evaluation method. Finally, the system was evaluated on several parameters such as accuracy, precision, recall, and so on