Skip to main content
. Author manuscript; available in PMC: 2015 May 1.
Published in final edited form as: Neuroimage. 2014 Jan 27;91:386–400. doi: 10.1016/j.neuroimage.2014.01.033

Figure 1.

Figure 1

Classification via matrix shrinkage and matrix completion. There are three main parts in this framework: feature selection, sample selection and matrix completion. Note that feature selection only involves training set. (Xs, Ys: Shrunk version of X and Y; Zc: Completed version of Z.)