Pengenalan Tutur Vokal Bahasa Indonesia Menggunakan Metode Multi Layer Perceptron

ABSTRACT: Indonesian language is a national language of Indonesia which can be interpreted as sound symbol that can be used as a communication device. In that communication process, information in the form of signal can be obtained. There are many researches on voice signal of Indonesian language. In this research, Indonesian vocal signal recognition automatization is developed by using Multilayer Perceptron to increase accuracy, sensitivity, specificity and precision. The recognition is done by using Discrete Wavelet Transform to 100 speakers that result in 500 characteristic data for training and testing. The classification process results inaccuracy as much as 95,9%, sensitivity 90,2%, specificity 97,5%, and precision 89,8%.
Keyword: Multi Layer Perceptron, Discrete Wavelet Transform, Vocal, Indonesian
Penulis: Risky Via Yuliantari, Risanuri Hidayat, Oyas Wahyunggoro, Anan Nugroho
Kode Jurnal: jptindustridd160356

Artikel Terkait :