Student's T-Hidden Markov Model for Unsupervised Learning Using Localized Feature Selection.
Yuhui Zheng,Byeungwoo Jeon,Le Sun,Jianwei Zhang,Hui Zhang
DOI: https://doi.org/10.1109/tcsvt.2017.2724940
IF: 5.859
2018-01-01
IEEE Transactions on Circuits and Systems for Video Technology
Abstract:Recently, the hidden Markov model (HMM) with student’s t-mixture model (SMM), called student’s t-HMM (SHMM) for short, has received much attention in unsupervised learning of sequential data. However, the current existing SHMMs fail to take into consideration of the relevant features embedded in local subspaces, thus influencing their performances in clustering. To address the problem, a novel SHMM is proposed by combining the measure of localized feature saliency (LFS) with SMM and utilizing two student’s t-distributions as subcomponents to respectively describe the distributions of useful features and non-salient “features,” with the purpose of accurately modeling the hidden state observation emission distributions of SHMM. Moreover, we exploit the variational Bayesian learning technique to simultaneously estimate the LFS, the number of components and other parameters of the herein proposed SHMM. Experimental results on both synthetic and real data sets demonstrate the improved robustness, effectiveness, and accuracy of our model.