Department of Statistics
University of Wisconsin-Madison
oem fits various penalized regression models using the Orthogonalizing EM algorithm. It is highly efficient for tall data. It provides estimation for penalties including the lasso, MCP, SCAD, elastic net, and group lasso.
This visualization allows for close inspection of a comparison of two estimating equations. One is the non-smooth Gehan-weighted estimating equation of the semiparametric accelerated failure time model and the other is a smooth approximation of it.