Speech-Based Number Recognition Using KNN and SVM

Speech-Based Number Recognition is a system that recognizes numbers based on the speech of the user. Most of the research makes use of English, Bangla, Tamil, etc., but the Malay language has received little attention. In this paper, the Malay numbers one through ten are recognized and implemented on devices consisting primarily of the Arduino UNO, the ELECHOUSE Voice Recognition Module v3, Microphone, and Light Emitting Diode. This system employs database creation, preprocessing, feature extraction, Mel-frequency cepstral coefficients, and classification utilizing using K-Nearest Neighbour and Support Vector Machine. Two experiments were carried out using 900 samples. In the first experiment, 80 percent of the training samples and 20 percent of the test samples were used. The second experiment utilized 70 percent of the training samples and 30 percent of the testing samples. The results show that the Support Vector Machine outperformed K-Nearest Neighbour with an average accuracy of 91.27 percent.