In this paper we propose a method which can infer causal links. In particular we focus on the diffculty of high dimensional cases. We begin by introducing a recent model that we will modify to improve performance. We add a skeleton and a Ridge penalty. We finish with some numerical simulations.