Automatic Vowel Recognition from Assamese Spoken words
P Sarma1, M Mitra2, M P Bhuyan3, V Deka4, S Sarmah5, S K Sarma6

1P Sarma*, Department of Information Technology, Gauhati University Guwahati, India. S Mitra, Department of Information Technology, Gauhati University Guwahati, India.
2M P Bhuyan, Department of Information Technology, Gauhati University Guwahati, India.
3V Deka, Department of Information Technology, Gauhati University Guwahati, India.
4S Sarmah, Department of Information Technology, Gauhati University Guwahati, India.
5S K Sarma, Department of Information Technology, Gauhati University Guwahati, India.

Manuscript received on 02 July 2019 | Revised Manuscript received on 09 July 2019 | Manuscript published on 30 August 2019 | PP: 2297-2304 | Volume-8 Issue-10, August 2019 | Retrieval Number: 10.35940/ijitee.J1301.0881019 | DOI: 10.35940/ijitee.J1301.0881019
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: Vowel plays the most important role in any speech processing work. In this research work, recognition of Assamese vowel from spoken Assamese words is explored. Assamese is a language which is spoken by major people in Brahmaputra Valley of Assam, Assam is a state which is situated in the North-East part of India. This automatic vowel recognition system is implemented by using three efficient techniques Support Vector Machine (SVM), K-Nearest Neighbor (KNN) and Random Forest (RF) classifier. The database used in the experiments is specially designed for this purpose. A list of phonetically vowel rich Assamese words is prepared for the experiment. As an initial effort, twenty different (20) words uttered by fifty-five (55) speakers are taken. Utterances from both male and female speakers are collected. Each utterance was repeated two times by every speaker. A database of the total of 2200 samples is prepared. After experimenting on different samples it is seen that Random Forest (RF) is giving the best performance compared to the other two classifiers. The performance of the system is shown with testing dataset and comparison is done. Outcome of this research work will enhance the Machine Translation from Assamese to any other language.
Keywords: LPC, Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Random Forest (RF).
Scope of the Article: Machine Learning