Bayesian Context Clustering Using Cross Validation for HMM-Based Speech Synthesis

Abstract This paper proposes a prior distribution determination technique using cross validation for HMM-based speech synthesis based on the Bayesian approach. The Bayesian method is a statistical technique for estimating reliable predictive distributions by marginalizing model parameters and its approximate version, the variational Bayesian method has been applied to HMM-based speech synthesis. Since prior distributions representing prior information about model parameters affect the model selection (e.g., decison tree based context clustering), the determination of prior distributions is an important problem. The proposed method can determine reliable prior distributions without tuning parameters and select an appropriate model structure dependently on the amount of training data. Key words Bayesian criterion, HMM-based speech synthesis, context clustering, prior distribution, cross valida-tio