<![CDATA[Stress is a feeling of emotional tension that can be caused by many things, such as work, studies, family, etcetera. If stress is not addressed quickly, it can worsen and have an impact on your health. Several studies have produced good results with a device that can tell someone's emotion by their voice. The purpose of this research is to know whether someone is experiencing stress and to know the level of stress using their voice. By using MFCC feature extraction and Artificial Neural Network machine learning is expected to know the level of stress using speech. This system works by using Raspberry Pi, which connects to a microphone. The Raspberry Pi will wait for a command that start and stop the recording from a buttnon, then when a voice is recorded it will extract its feature and predict the result. That result will be shown on LCD. In this research, the dataset that is being used is Speech Under Simulated and Actual Stress (SUSAS) dataset, which contains 1860 utterances. The result of this research using 30 samples from the dataset is 90% accuracy, but when building the artificial neural network model, the testing accuracy is only 76%. where the average computing time while testing the system is 2.65 seconds.]]>
Copyrights © 2022
