Unknown noise estimation with DPM model and its application
Guo Shuxia, Bo Yang, Ning Liu · 2010
Dirichlet process mixture (DPM) model, which is the state-of-the-art Bayesian nonparametric model in Statistics, was introduced here. We explored its ability to model and estimate nongaussian unknown noise and applied it to dynamic estimation problem. The algorithm was given and a real world example of GPS estimation problem demonstrated the efficiency of our algorithm.