Application of Feature Transformation and Learning Methods in Phoneme Classification

TitleApplication of Feature Transformation and Learning Methods in Phoneme Classification
Publication TypeConference Paper
Year of Publication2001
AuthorsKocsor A, Tóth L, Felföldi L
Conference NameEngineering of Intelligent Systems : 14th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, IEA/AIE 2001, LNAI vol. 2070
Date PublishedJune
PublisherSpringer-Verlag GmbH
Place PublishedBudapest, Hungary

This paper examines the applicability of some learning techniques to the classification of phonemes. The methods tested were artificial neural nets (ANN), support vector machines (SVM) and Gaussian mixture modeling. We compare these methods with a traditional hidden Markov phoneme model (HMM) working with the linear prediction-based cepstral coefficient features (LPCC). We also tried to combine the learners with feature transformation methods, like linear discriminant analysis (LDA), principal component analysis (PCA) and independent component analysis (ICA). We found that the discriminative learners can attain the efficiency of the HMM, and after LDA they can attain practically the same score on only 27 features. PCA and ICA proved ineffective, apparently because of the discrete cosine transform inherent in LPCC.