Data dependent asymmetric kernels for estimating the density function (Q2023834)

In this paper, the authors propose a new smooth density estimator which uses data dependent asymmetric kernels generated by an unimodal symmetric kernel and study its asymptotic properties. Simulation studies and data analysis indicate that the proposed estimator performs better compared to more common estimators based on a symmetric kernel. Multivariate generalizations of the method are also considered with specific attention to the bivariate case.