<![CDATA[Speech is a part of the human that has unique characteristics so that it can be distinguished from one person with someone else. Speech delivered, has a variety of information so that in its application it can be used to carry out voice commands using speech. In signal processing, Mel Frequency Cepstrum Coefficient (MFCC) is a method used for feature extraction. In this study, MFCC is used as a feature extraction method using Matlab R2017a and K-Nearest Neighbor (KNN) software used to identify and classify voice commands spoken by the speaker using speech pattern patterns obtained from the MFCC. This study uses 10 training data for each voice command word consisting of open, close, message and gallery, and 5 test data for each voice command word. Voice data is used using different words and different speakers. This research yields an accuracy level of 60% in voice Buka, 60% in voice Tutup, 60% in voice Pesan and 65% in voice Galeri.]]>
Copyrights © 2026
