Some Robust Density Estimators for Spatial Point Processes

A small number of robust, distance-based estimators of density for spatial point processes have been proposed in recent years. A method for producing robust estimators and improving on existing ones is suggested. The method attempts to modify existing estimators of density by a factor that counterbalances bias, the factor qualitatively varying with type of spatial pattern. Simulation results and application to forestry data show that the proposed estimators work well and that improvements can be made to what is perhaps the best estimator of density to date.