Gaussian Estimation


Gaussian density estimation is supported by Scatterplots.  It can be accessed in the graphic menu in the entries Marron /arrows and Marron / sig. arrows.

This density estimation is calculated in the following way:

  1. Binning of the data
  2. Calculate density estimation with binned data and a gaussian kernel function
  3. Calculate the gradient (the change of the density estimation) for every bin of the data
  4. Calculate significant gradients

If Marron / arrows was chosen, the last step will be skipped and every gradient will be painted into the graph.

See the following examples:

 

Using ALT-UP/ALT-DOWN it is possible to increase/decrease the binsize in step 1.

Using ALT-LEFT/ALT-RIGHT it is possible to increase/decrease the window width (a factor of the binsize) in 2.