Speech Recognition Using Supervised and Unsupervised Learning Techniques
Amber Singh, R. S. Anand · 2015
This paper presents an investigation of speech recognition accuracy of unknown test patterns using five models when classes for classification of unknown test patterns increase from three to five. GMM, SVM, MLP, RBPNN and LVQ are used and their speech recognition accuracy are studied on five isolated digits. In the first experiment, three isolated words from zero to two are used for training and testing. In the second experiment, five isolated words from zero to four are used. Finally classification accuracy of each classifier is found using unknown test patterns and conclusion is made.