|
Abstract:
|
We consider the problem of performing matrix completion with side information on row-by-row similarities and column-by-column similarities. We build upon recent proposals for matrix estimation with smoothness constraints with respect to row and column graphs. In this talk, we address the unaddressed issue in model selection in these approaches, namely how to choose an appropriate amount row and column smoothing. We also discuss how to exploit the sparsity structure of the problem to scale up the estimation and model selection procedure. We present simulation results and an application to predicting associations in high-dimensional imaging-genomics studies.
|
