Model parameter estimation for mixture density polynomial segment models

In this paper, we propose parameter estimation techniques for mixture density polynomial segment models (MDPSMs) where their trajectories are specified with an arbitrary regression order. MDPSM parameters can be trained in one of three different ways: (1) segment clustering; (2) expectation maximization (EM) training of mean trajectories; and (3) EM training of mean and variance trajectories. These parameter estimation methods were evaluated in TIMIT vowel classification experiments. The experimental results showed that modelling both the mean and variance trajectories is consistently superior to modelling only the mean trajectory. We also found that modelling both trajectories results in significant improvements over the conventional HMM. © 1998 Academic Press

• Type: Journal
• Year: 1998
• Where: CSL
• Authors: Toshiaki Fukada, Kuldip K. Paliwal, Yoshinori Sagisaka
• Keywords: Automated Reasoning, CSL 1998, Parameter Estimation, Trajectories, Variance Trajectories