Learning with Segment Boundaries for Hierarchical HMMs
Hierarchical hidden Markov models (HHMMs) can be used for time series segmentation. However, it is difficult to obtain a desirable segmentation result, because the form of learning for HHMMs is unsupervised. In the paper, we present a semisupervised learning algorithm for HHMMs. It is semisupervised in the sense that the supervisor teaches segmentation boundaries but not segment labels. The learning performance of the proposed algorithm is demonstrated through an experiment using music data.
KeywordsHide Markov Model Segmentation Result Segment Boundary Horizontal Transition Hide Markov Model Parameter
Unable to display preview. Download preview PDF.
- 2.Cohen, I., Garg, A., Huang, T.S.: Emotion recognition from facial expressions using multilevel HMM. In: Neural Information Processing Systems Workshop on Affective Computing, Denver, USA (2000)Google Scholar
- 3.Murphy, K.P., Paskin, M.A.: Hierarchical HMMs, Technical reports (2002)Google Scholar
- 4.Rabiner, L., Juang, B.H.: Fundamentals of Speech Recognition. Prentice Hall PTR., Englewood Cliffs (1993)Google Scholar