Speech Recognization Based on Support Vector Machine
Aiming at the deficiency of the local minimum occurring in neural network used for speech recognition, the paper employs support vector machine (SVM) to recognize the speech signal with four different components. First, SVM is utilized to perform the speech recognition. Then, the results are compared with those obtained by the BP neural network method. The comparison shows that SVM effectively overcomes the local minimum existing in neural network and has the advantages of the accurate and fast classification, indicating that SVM looks feasible to recognize the speech signal.
Cai Suo Zhang
L. Zhang, "Speech Recognization Based on Support Vector Machine", Advanced Materials Research, Vols. 433-440, pp. 7516-7521, 2012