In alt.ufo.reports MrPostingRobot@kymhorsell.com wrote:

There are 3 types of volcano.
EXECUTIVE SUMMARY:
- Similar to a prev study, we look at the distribution of UFO sightings
 from the major ~50 volcanoes across the USA.
- Surprisingly the volcanoes break down into 3 groups.  (a) Those have
 have "no effect". (b) Those that attract UFO sightings with a power
 law around r^(-.25). (c) Those that repel UFO sightings with a
 power law around sqrt(r).
- We've seen the sqrt(r) law before with respect to US AFB.  It can be
 interpreted to mean someone is wary of something flying under
 continuous acceleration (transit time proportional to sqrt(r)) like
 a missile. Or another UFO.
- The power law r^(-.25) is a bit of a puzzle. It doesn't exactly jive
 with some volcanoes being "home bases" for UFO activity.  That
 would be r^(-2). It may suggest UFO's have a flying pattern that
 weights longer distances approximately r^1.75.  Why that pattern is
 not obvious. But if you have access to machinery that can
 super-cruise all day long you probably don't use much it to cruise
 your own street.

...

While I'm here I'll add this addendum.

I had an unusual visit today. Not only does it take some gumption to
drive 50 mi into Australia's "virus central" region to visit,
but the visitor was a mathematician and also not totally dismissive
of UFO's, esp in light of recent publications in TIME which he seems
to read. :)

We were talking about this mathematical model that seems to suggest
some volcanoes seem to "repel" UFO sightings.

He immed suspected the modeling s/w had just latched onto
one group of volcanoes, that happen to mostly be off the main AK coast,
and assigned them a special property and the idea they were
different from the US48 volcanoes that seemed to have a different
"law of attraction".

So I had to cook up something fast. And the first thing worked.
Running the model backwards you can ask according to the UFO sightings
data where would volcanoes that seem to repel sightings be?
The sightings data scattered over the US landscape is like charges creating
an electric field, where in the field are the repellant bits?

So running the model backwards we plotted out where the repellant
volcanoes would most likely be.

They all turned out to be around Alaska.

So it's not that the AI simply got distracted by the latitude
of some volcanoes and assigned them a special property by mistake,
the whole region up there has this property. Plus a couple of other odd
cases in Oregon and Hawaii.


The plot is here <kymhorsell.com/UFO/tests.gif>.
The green crosses are the locations that "should" have a UFO-repellant
volcano according to how the total distribution of UFO sightings
lays, and the blue cirles are where they are actually located.

Reasonable match.