|
|
|
| Study on Hybrid Sampling Inference for Dirichlet ProcessMixture of Gaussian Process Model |
| LEI Juyang1，2，HUANG Ke1，XU Haixiang3，SHI Xizhi1 |
| (1.State Key Laboratory of Mechanical System & Vibration, Shanghai Jiaotong University,Shanghai 200240, China； 2.College Mechanical Engineering, Shanghai University of Engineering Science, Shanghai 201620, China; 3.Special Equipment Inspection Institute, Shanghai 200062, China) |
|
|
|
|
Abstract Dirichlet process mixture of Gaussian process model was proposed to reveal the intrinsic mechanism of multimodel of complex dynamic system architecture data. As for the difference between the mean structure and covariance structure of sparsity, parametric a priori and nonparametric a priori were designed based on the hybrid sampling framework of Polya urn sampling and overrelaxed sliced sampling. The hybrid sampling will not only be implemented under the unified MetropolisHasting probability evaluation criteria , but also be able to overcome the shortcomings of Gaussian random walk. Markov chain samples can be quickly and easily extended. The simulation results show that the hybrid sampling algorithm has more extensive adaptabilities and more accurately predictive effect than that of Gaussian process regression model and GPFR mixture model.
|
|
Received: 02 April 2009
Published: 26 February 2010
|
|
|
|
|
|
| [1] |
WANG Xiaoxia,YANG Fengbao,LIN Suzhen,SHI Dongmei. A Method of Risk Assessment Based on Flexible Similarity Measurement and Possibility Skewness[J]. J. Shanghai Jiaotong Univ.(Sci.) , 2014, 48(10): 1440-1445. |
| [2] |
WU Wentong1,LI Yuanxiang1，Wei Banghe2,Zheng Silong1. Incremental ISOMAP Method Based on Locally Estimated Geodesic Distance[J]. J. Shanghai Jiaotong Univ.(Sci.) , 2013, 47(07): 1082-1087. |
| [3] |
LEI Ju- Yang. Speech Enhancement Based on Chinese Restaurant Process[J]. J. Shanghai Jiaotong Univ.(Sci.) , 2013, 47(04): 635-639. |
|
|
|
|
