Semiparametric Gaussian Random Fields
Inspired by the previous work in Gaussian graphical models as well as nonparametric density estimation, we are interested in combining these two areas together. By estimating part of the data nonparametrically, we are able to relax the distributional assumptions and allow higher data dimension.
First-stage estimation have been completed, currently we are working on computational optimazation and theoretical proof.
Source: Sparse Gaussian Conditional Random Fields: Algorithms, Theory, and Application to Energy Forecasting. Wytock et.al. (2013)