PENDETEKSIAN KEBOHONGAN MENGGUNAKAN ALGORITMA VITERBI DAN TRANSFORMASI HANKEL PADA SUARA SECARA REAL TIME
Keywords:Viterbi, Hankel, Fourier, True detection
The lie detection system on the sound can be detected using the Viterbi algorithm. In a previous study stated that by using Viterbi the results of true detection were closer to the optimal results but there were still some data that resulted in low values and the sound samples that were tested were still using previous data records. In this study will use the Hankel transformation which is known as the Fourier-Bessel transformation. Just as the Fourier transform for an infinite interval is related to the Fourier series during a finite interval, so Hankel's transformation of an infinite interval is related to the Fourier-Bessel series during a finite interval. This research carried out the process of recognizing and training lying sound samples on the words go, walk and move then tested by simulating them on the training data and test data to produce the percentage of recognition and classification of the lie sounds. The data tested in this study used samples that were tested in real time. The results showed that the lie detection system using the Viterbi algorithm had a true detection range of 89% in the word go, 87% in the word walk and 84% in the word move. Whereas using Hankel transformation has a true detection range of 94% in the word go, 94% in the word walk and 95% in the word movemensimulasikannya pada data latih dan data uji untuk menghasilkan persentase pengenalan dan klasifikasi suara kebohongan. Data yang diuji dalam penelitian ini menggunakan sampel yang diuji secara real time. Hasil penelitian menunjukkan bahwa sistem deteksi kebohongan dengan algoritma Viterbi memiliki jangkauan deteksi benar sebesar 89% pada kata go, 87% pada kata walk dan 84% pada kata move. Sedangkan penggunaan transformasi Hankel memiliki jangkauan deteksi sebenarnya 94% pada kata go, 94% pada kata walk dan 95% pada kata move.