Universiti Teknologi Malaysia Institutional Repository

Malay speaker dependent digits recognition with improved backpropagation

Mohamad, Ummu Salamah and Mahmod, Ramlan and Shamsuddin, Siti Mariyam (2004) Malay speaker dependent digits recognition with improved backpropagation. Jurnal Teknologi Maklumat dan Multimedia, 1 . pp. 1-14. ISSN 1823-0113

[img] PDF - Published Version
643Kb

Abstract

This paper presents a study of a Malay speaker dependent recognition using improved Neural Network (NN). The performances are evaluated for recognition of the isolated Malay digits of "0" through "9". The Error Backpropagation (BP) and an improved error signal of the BP are used in this study. Experiments are carried out by comparing the recognition rates and convergence time of the standard BP and improved BP, as well as the effects of normalisation techniques on Malay speaker dependent data. The utterances are represented using the Linear Prediction Coding (LPC) method. The results show that the improved BP outperforms the standard BP in terms of its convergence with better recognition rates for unnormalised data. For the effects of normalisation data, the unit simple method gives better result compared to unit range and unit variance with improved BP gives faster convergence and higher recognition rates. ABSTRAK: Makalah ini membincangkan kajian pengecaman jurucakap Melayu tidak bebas dengan menggunakan rangkaian neural pembaikan. Prestasi dinilai untuk pengecaman digit terpisah Melayu '0' hingga '9'. Rambatan balik dan isyaratralat pembaikan telah digunakan dalam kajian ini. Ujikaji telah dijalankan dengan membandingkan kadar pengecaman dan masa penumpuan rambatan balik piawai, rambatan balik pembaikan dan kesan teknik-teknik penormalan pada datajurucakap tidak bebas Melayu. Ujaran-ujaran diwakili dengan pengkodan ramalan lelurus. Keputusan menunjukkan bahawa, rambatan balik pembaikan mengatasi rambatan balik piawai dari sudut penumpuan dengan kadar pengecaman yang lebih baik untuk data yang tidak dinormalkan. Bagi kesan penormalan data, kaedah mudah memberikan keputusan baik berbanding dengan unit julat dan unit varians dengan rambatan balik pembaikan memberikan penumpuan lebih cepat dan kadar pengecaman yang lebih tinggi

Item Type:Article
Uncontrolled Keywords:Malay language, Spoken Malay, Neural networks techniques, audio data
Subjects:Q Science > QA Mathematics > QA75 Electronic computers. Computer science
ID Code:196
Deposited By: Kamariah Mohamed Jong
Deposited On:12 Nov 2012 01:03
Last Modified:12 Nov 2012 01:27

Repository Staff Only: item control page