4
$\begingroup$

I have been doing a bunch of spherical statistics using Stereonet and there is a function for producing contour plots of data density.

One of the options in this function is a choice between Kamb contours and 1% area contours. What is the difference between these two and what are the benefits/disadvantages of each?

The main difference seems to be that 1% area contours fit much more tightly to the data.

Kamb contours:

enter image description here

1% area contours:

enter image description here

$\endgroup$
  • 1
    $\begingroup$ Neither term "Kamb" or "1% area" is common in statistics. It's possible, by searching the peer-reviewed literature, to determine what a "Kamb contour" is (while, at the same time, establishing that many people who describe these haven't any clear or accurate idea of what they are!), but it's harder to find an adequate definition of "1% area contours." Would you happen to be able to give us such a definition or a reference to one? $\endgroup$ – whuber Jan 11 '18 at 15:56
  • 1
    $\begingroup$ @whuber I did find this but I'm still not entirely sure what the difference between the two is. $\endgroup$ – bon Jan 11 '18 at 17:44
  • $\begingroup$ The pieces of that book that are shown suggest it contains an accurate account of the Kamb (1959) method. (It's a uniform KDE on the sphere or projective plane that uses a particular criterion to determine the kernel radius.) The challenge is to find a similar account of whatever this "1% area" method might be. It's possible to guess based on the context, the name, the purpose, and your examples, but it would be nice to have some authority to rely on rather than guessing. $\endgroup$ – whuber Jan 11 '18 at 18:22
  • $\begingroup$ @whuber What is your guess of what it is? $\endgroup$ – bon Jan 12 '18 at 9:26
  • $\begingroup$ My best guess is that both of these are uniform spherical KDEs. They differ only by kernel radius. The Kamb radius is chosen to limit the standard error of the mean within any randomly located circle (and evidently is relatively large in this example) while the "1% area" radius is literally the one whose circle is a specified proportion of the total area of the hemisphere, by default 1%. $\endgroup$ – whuber Jan 12 '18 at 15:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.