| Abstrakt: |
This research is addressed to modern hospitals with nurses call innovation invention to solve ordinary ways. However, this equipment is still not optimal because the patient's family can only make calls to nurses without being able to deliver the assistance they want, a research entitled "Vocal Letters Classification with Audio Processing for the Nurse Room Notification System with the Mel Frequency Cepstrum Coefficient and Backpropagation Method" Neural Network." This study uses the microcontroller system divided into two parts, namely hardware design and software design. This study's results are a speech recognition nurse room system based on speech recognition as a means of communication between patients and caregivers in hospital inpatient rooms. The results of this study are divided into two tests. First data training vocal a: 100%, I: 80%, U: 100%, E: 100%, O: 80% so that it obtained a total accuracy of 92%. The second test results of vocal test data: 100%, I: 100%, U: 0%, E: 100%, and O: 10% so that the total accuracy is 80%. The results of both tests concluded that the pattern could be recognized quite accurately. [ABSTRACT FROM AUTHOR] |
