Quadratic approximation for nonconvex penalized estimations with a diverging number of parameters

We propose an approximated penalized estimator (APE) that covers various statistical models and nonconvex penalties including the smoothly clipped absolute deviation (SCAD) penalty (Fan and Li, 2001) as a special case. The APE achieves the oracle property with a diverging number of parameters which extends the results of Kwon et al. (2011). Several numerical studies confirm the theoretical results.