DescriptionInvited speaker discussing the paper - Two-layer Mixture of Gaussian Processes for Curve Clustering and Prediction.
The mixture of Gaussian processes is capable of learning any general stochastic process based on a given set of (sample) curves for regression and prediction. However, it is ineffective for curve clustering analysis and prediction when the sample curves come from different stochastic processes as independent sources linearly mixed together. In fact, curve clustering analysis is a very challenging problem in the modern big data era. Recently, we have established a two-layer mixture model of Gaussian processes to describe such a mixture of general stochastic processes or independent sources, especially for curve clustering analysis and prediction. This talk describes the learning paradigm of this new two-layer mixture of Gaussian processes, introduces its MCMC EM algorithm and presents certain effective practical applications on curve clustering analysis and prediction.
|Period||10 Jan 2020|
|Visiting from||Peking University (China)|
|Degree of Recognition||International|