Projection estimates of point processes boundaries (Q1400118)

The authors study the problem of estimating a bounded set \(S\) of \(\mathbb{R}^2\) given a finite random set \(\Sigma\) of points drawn from the interior. They consider the special case where \(\Sigma\) is the set of points of a homogeneous Poisson process \(N\) whose support is \(S=\{(x,y)\in \mathbb{R}^2\mid 0\leq x\leq 1,\;0\leq y\leq f(x)\}\), where \(f\) is an unknown function. The estimation of \(S\) reduces to the estimation of the function \(f\). The estimator is based both on projections on \(C^1\) bases and on extreme points of the process \(N\). Conditions on the Dirichlet kernel associated with the \(C^1\) bases for various kinds of convergence and asymptotic normality are established. A method for reducing the bias is also proposed.