A Comparison of the LBG, LVQ, MLP, SOM and GMM Algorithms for Vector Quantisation and CLustering Analysis
Roberto B. Togneri, Dariush Farrokhi, Yaxin Zhang, Yianni Attikiouzel · 1992
We compare the performance of five algorithms for vector quantisation and clustering analysis: the Self-Organising Map (SOM) and Learning Vector Quantization (LVQ) algorithms of Kohonen, the Linde-Buzo-Gray (LBG) algorithm, the MultiLayer Perceptron (MLP) and the GMM/EM algorithm for Gaussian Mixture Models (GMM). We propose that the GMM/EM provides a better representation of the speech space and demonstrate this by comparing the GMM with the LBG, LVQ, MLP and SOM algorithms in phoneme classification and digit recognition. INTRODUCTION Currently, the most popular approach to speech recognition is the combination of Vector Quantization (VQ) for the encoding of segments of speech with a Hidden Markov Model (HMM) for the classification of sequences of segments. The VQ stage is usually unsupervised since no a priori assumptions on the speech class distribution is made. The VQ algorithm proposed by Linde et al. (1980) , and subsequently referred to as the LBG algorithm is currently the algo...