Using Parzen Window approach

Question

When is preferable using Parzen Windows approach, so a nonparametric approach, instead to a parametric approach?

Answer 1

It's under kernel density estimation (KDE). It's non-parametric so you don't have to assume a form. Otherwise, you'd need to assume some form of distribution, (e.g. exponential, normal, mixture of normals etc) and estimate the parameters. With Parzen, you could fit a distribution without making any initial assumptions.