Probability density estimation in astronomy

In this paper we consider three empirical estimators (Kernel, adaptive Kernel and parametrizing families of Johnson) of the probability density function, which are alternatives to the classical Histogram. By means of numerical simulations the superiority of such estimators is shown in the context of some typical astronomical problems. Given the simplicity of implementation and the efficiency of calculation, the use of these methods is recommended. Indication for the implementation is provided.