|
Abstract:
|
In this article, we propose a new and simple method for estimating coefficient of the direction in single index model in the large $p$ small $n$ settings. Our proposed method enables existing approaches for $n>p$ to be adapted to the $n< p$ problems. Hence, we could obtain an estimate even if $n< p$. In addition, the method can successfully select important predictors. We provide theoretical proof for our algorithm. Simulations and real data analysis demonstrate the advantages of our method.
|
