Kernel Density Estimation KNN Question

[Maikelmoore]

Hello,

As before posted (pysal/pointpats#67), we are having a second question regarding pysal during writing our master thesis:

About Kernel Density Estimation: Therefore we are using the Kernel calculation function in your scripts „weights.py“ „distance.py“ (libpysal, version 4.3.0, which we attach to this email, lines 597 - 620) to get the Kernel values which we then plug into the KDE formula. When it comes to the choice of band widths, we see the options of a fixed and non-fixed number of neighbours (K) and that the default number is set to 2.
The question which arose is: Is there a special reason behind setting the default number of neighbours to 2? Is there a way to get an optimal K for such cases?

Furthermore, are there any ways to calculate a bandwidth accfilesording to Silvermans or Scotts Rule in pysal?

Thank you in advance for your help.

weights.txt
distance.txt