Fit a non parametric copula

In this example we are going to estimate a Normal copula from a sample using non parametric representations.

import openturns as ot
import openturns.viewer as otv

Create data

R = ot.CorrelationMatrix(2)
R[1, 0] = 0.4
copula = ot.NormalCopula(R)
sample = copula.getSample(30)

Estimate a Normal copula using BernsteinCopulaFactory

distribution = ot.BernsteinCopulaFactory().build(sample)

Draw fitted distribution

graph = distribution.drawPDF()
view = otv.View(graph)

Estimate a Normal copula using KernelSmoothing

distribution = ot.KernelSmoothing().build(sample).getCopula()
graph = distribution.drawPDF()
view = otv.View(graph)

Display all figures

otv.View.ShowAll()