ALGORITHM FOR CLUSTERING CONTINUOUS DENSITY HMM BY RECOGNITION ERROR

This paper presents a clustering algorithm producing multiple whole-wo rd continuous density hidden Markov models (CDHMM) for isolated word r ecognition systems. The algorithm estimates a minimum number of CDHMM per word that approaches or satisfies a minimum predefined word-depend ent recognition accuracy in the training set. Significantly lower memo ry requirements and a better and more uniformly distributed recognitio n accuracy among the words of the vocabulary are measured by comparing this algorithm with the modified K-means clustering method.