A gentle introduction to kernel density estimation

1 · Stathis Kamperis · Dec. 8, 2020, midnight
Introduction Suppose we have some low dimensional data (1 or 2 variables). How do we start exploring them? Usually, one of the first steps is to plot their histogram to get a feeling of how they are distributed. Histograms are nice because they provide a fast and unambiguous way to visualize our data’s probability distribution. However, they are discrete, and sometimes it is useful to have a smooth estimation of the underlying probability density function (PDF) at hand. There are two ways to ...