Title: Wavelet-probabilistic neural network-based intelligent system approach for spoken Hindi paired word classification

Authors: Dinesh Kumar Rajoriya, R.S. Anand, R.P. Maheshwari

Addresses: Department of Electrical Engineering, Indian Institute of Technology Roorkee, Roorkee-247667, Uttarakhand, India. ' Department of Electrical Engineering, Indian Institute of Technology Roorkee, Roorkee-247667, Uttarakhand, India. ' Department of Electrical Engineering, Indian Institute of Technology Roorkee, Roorkee-247667, Uttarakhand, India

Abstract: This paper has proposed an idea to develop a speech controlled authenticated voice enabled system based on hybrid approach consisting of wavelet and probabilistic neural network (PNN) techniques. This hybrid approach is endeavoured for spoken paired word classification that is conceived of as an integral part of a speaker dependent intelligent system. Five hundred templates of spoken Hindi (Indian national language) paired word (SHPW) have been used for experimental validation of the proposed approach. The SHPW signal attributes are extracted by wavelet technique and classification/recognition has been examined with the help of PNN. Correct classification rate of proposed speaker dependent SHPW intelligent system is found to be 100% on considered database.

Keywords: intelligent system design; spoken paired word; SPW; wavelet transform; spoken Hindi paired word; SHPW classification; probabilistic neural networks; PNNs; correct classification rate; CCR.

DOI: 10.1504/IJICA.2011.041916

International Journal of Innovative Computing and Applications, 2011 Vol.3 No.3, pp.152 - 159

Received: 15 Aug 2010
Accepted: 15 Mar 2011
Published online: 21 Mar 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article