文档介绍：puting 51 (2003) 225–235
ate/
Cooperative supervised and unsupervised learning
algorithm for phoneme recognition in continuous
speech and speaker-independent context
Najet Arous∗, Noureddine Ellouze
Uniteà de Recherche: Signal, Image, Reconnaissance de Formes, Groupe: Reconnaissance Vocale,
Ecole Nationale d’Ingenieursà de Tunis, BP-37 Campus Univesitaire 1002 Tunis, Tunisia
Received 5 July 2001; accepted 8 May 2002
Abstract
works have been traditionally considered as an alternative approach to pattern recog-
nition in general, and speech recognition in particular. There have been much ess in practical
pattern recognition applications using works including multi-layer perceptrons, radial
basis functions, and anizing maps (SOMs).
In this paper, we propose a system of SOMs based on the association of some supervised and
unsupervised learning algorithms inherited from the most popular workin the unsu-
pervised learning category, SOM. The case study of the proposed system of SOMs is phoneme
recognition in continuous speech and speaker independent context. Also, we propose a way
to save more information during training phase of a Kohonen map in the objective to amelio-
rate speech recognition accuracy. The applied SOM variants serve as tools for developing intel-
ligent systems and pursuing artiÿcial intelligence applications.
c 2002 Elsevier Science . All rights reserved.
Keywords: work; Supervised learning; Unsupervised learning; anizing map; Continuous
speech recognition
1. Introduction
Among the most challenging areas of artiÿcial intelligence and pattern recognition,
automatic speech recognition is one of the di8erent topics which has attracted particular
∗ Corresponding author.
E-mail addresses: najet.******@ (N. Arous), n.******@ (N. Ellouze).
0925-2312/03/$ - see front matter c 2002 Elsevier Science . All rights reserved.
PII: S0925-2312(02)00618-5
226 N. Arous, N. Ellouze / puting 51 (2003)