Signed variable optimal kernel for non-parametric density estimation
arXiv:2102.07858
Abstract
We derive the optimal signed variable in general case kernels for the classical statistic density estimation, which are some generalization of the famous Epanechnikov's ones.