By Topic

Speech Recognition Method Based on Linear Descending Inertia Weight PSO Algorithm Optimizing SVM Kernel Parameters

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Jing Bai ; Coll. of Inf. Eng., Taiyuan Univ. of Technol. Taiyuan, Taiyuan, China ; Yueling Guo

An important factor that influences the performance of support vector machine is how to select the parameters. Particle swarm optimization is an efficient algorithm and it is broadly used in many research areas like pattern recognition. In order to improve the learning and generalization ability of support vector machine and enhance the speech recognition system accuracy, a method for searching the Gaussian kernel support vector machine optimal parameters (C,¿) based on particle swarm optimization is proposed and a speech recognition system based on support vector machine using the optimal parameters is constructed in this paper. The inertia weight, a crucial parameter of the particle swarm optimization, is adopted in linear descending adjusting method. The speech data is isolated, non-specific and middle vocabulary words. The speech feature we used is modified MFCC feature. Experiments indicate that it is an efficient approach for parameters selection of support vector machine and has higher correct speech recognition rates than default parameters of the support vector machine open source software.

Published in:

Natural Computation, 2009. ICNC '09. Fifth International Conference on  (Volume:1 )

Date of Conference:

14-16 Aug. 2009