XPost: alt.ufo.reports   
      
   EXECUTIVE SUMMARY:   
   - The AI's have been busy extracting data from UK missing persons   
     police reports.   
   - We check several random but representative regions from the ones   
     available to see whether there are similarities with the previous   
     look-see at FBI missing persons data. There are. There are also some   
     differences.   
   - Most UK regions show a strong +ve correlation between local UFO   
     sightings and missing persons. One of the 4 regions selected shows   
     a strong -ve corr; people are "re-appearing" in Manchester in   
     spooky statistical parallel with UK UFO sightings.   
   - US UFO sightings have a weak but statistically relevant corr with UK   
     missing persons -- but all are -ve. IOW more US sightings relate to   
     fewer UK missing persons; more UK sightings relate to more UK   
     missing persons. It seems the black triangles may be stretched and   
     operate on only one continent at a time!   
      
      
   Let's recall a prev post showing some statistical links between UFO   
   sightings and FBI missing person data.   
      
   It seems certain categories of missing persons have a suspiciously   
   similar "shape" to the month-by-month meanderings of UFO sightings.   
   Other groups show no statistical link with UFO sightings.  The   
   patterns seemed consistent with the folklore that says perhaps "many"   
   people are abducted by persons unknown for reasons unknown but   
   connected with UFOs.   
      
   The AIs have been scrounging around for similar data from other   
   jurisdictions to compare with the patterns in the FBI data.   
      
   They have now managed to get detailed reports for 4 years that --   
   unlike the FBI reports previously analyzed -- break down missing   
   persons cases by region. The AI's are interested in having data broken   
   down geographically because it allows them to compare the trends   
   between regions with a slew of demographics they have on geographic regions.   
      
   In this case if they can establish regions with certain attributes are   
   more likely and others are less likely to see UFO-related missing   
   persons it could establish some kind of "reason" people are going   
   missing in patterns suspiciously similar to the wax and wain of UFO activity.   
      
   In the prelim decrypting of PDF's the AI's have selected several years   
   and several regions across England to establish some baselines.   
   Further data scrounging will aim to fill in more years with the same   
   kind of data to validate (i.e. replicate) these patterns.  It's   
   standard practice to use part of the data to "estimate parameters" and   
   "unseen data" later to re-estimate the same parameters and see if they   
   are statistically identical to the first set. If so -- you have something. :)   
      
   The selection includes a couple of larger cities -- Manchester and   
   London -- and a couple of "rural" areas.   
      
   Interestingly, 3 of the 4 show identical patterns that further reveal   
   a possible pattern of UFO activity across the N Hem.   
      
   As usual the basic tools involve a time series regression of UFO   
   activity (at the granularity of months this time) to determine whether   
   the "waveform" of monthly UFO sightings is statistically similar to   
   the waveform of quarterly missing persons data from the selected regions.   
      
   To spice things up 2 subsets of the UFO data were selected -- the   
   overall monthly counts from NUFORC and those sightings just associated   
   with the UK. The NUFORC accepts sightings from around the world on its   
   web report form and they do receive a "representative sample" from   
   dozens of countries. While analysis shows the number of "outside N   
   Am" data is very small compared with specialised country databases   
   e.g. GEIPAN (France), and even aggregate data from Australia's   
   various state-based UFO groups, it does seem statistically relevant   
   and highly correlates with national data obtained elsewhere.   
      
   The summary table is thus:   
      
   Sightings	UK region		R2	    Beta       95% CI   
   nuforcuk	metro 			0.95027414  129.292    21.3032   
   nuforcuk	devon.and.cornwall 	0.77594943  3.92015    1.20748   
   vadj		northyorkshire 		0.63070176  -0.0833505 0.0351008   
   vadj		devon.and.cornwall 	0.53745547  -0.196032  0.0597227   
   nuforcuk	northyorkshire 		0.43959080  3.21537    1.60521   
   nuforcuk	manchester 		0.43945404  -12.4379   7.18815   
   vadj		manchester 		0.34546658  -0.386339  0.205858   
   vadj		metro 			0.28905595  -0.394062  0.215392   
      
   The table is sorted by R2 ("explanation power").   
      
   "vadj" is the codename of monthly NUFORC sightings data.  "nuforcuk"   
   is just the sightings that mention "UK" as either the location or   
   appears in the body of the UFO report.   
      
   The "best model" relates UK sightings with missing persons in metro   
   London.  Each UK UFO sighting over the period 2009-2011 is associated   
   with 129+-21 missing Londoners. The month-to-month variation in UK   
   sightings tracks 95% of the month-to-month (the s/w converts the   
   quarterly reports into monthly data by just dividing by 3 -- a   
   conservative method that should, if anything, reduce the R2) variation   
   in missing persons. An incredible correlation.   
      
   of course we can't prove what the reason for the correlation is.  Maybe   
   a Londoner hears a story about a UFO seen in the previous few days and   
   packs their bags and moves to Europe.   
      
   And of course there is the more direct causal connection: someone got   
   beamed up by a black triangle.   
      
   At the other end of the spectrum is the association between overall N   
   Am UFO sightings and missing Londoners.  That model explains only   
   ~29% of the missing persons data.  And the \beta is -ve. Somehow more   
   UFO sightings in N Am are associated with fewer missing Londoners. The   
   AIs say this is unusual since missing person data shows distinct   
   seasonal trends and the UK and US have similar seasonal   
   patterns. We'll return to this point, below.   
      
   Of the 4 regions selected for this initial run Devon&Cornwall comes   
   next as a close connection. UK UFO sightings track about 77% of   
   missing persons in the region. The \beta is +ve. For each UK sighting   
   there is around 4 additional missing persons in the same month.   
      
   Again, when we look at how the region relates to total UFO sightings   
   (i.e. mostly from N Am) we again find a -ve beta but a good solid link   
   (R2).  It seems now in 2 cases we have found "US" sightings are -ve   
   related to UK missing people; but UK sightings are +ve related.   
      
   So it sounds like something is either going on in N Am or is going on   
   in the UK.  "Whatever" oscillates between the US and UK.  At some   
   times the "whatever" is in the US and people are disappearing there in   
   response. At some times the "whatever" is in the UK and people are   
   disappearing there.   
      
   The "whatever" sounds a bit stretched and seemingly can't operate in 2   
   well separated places at the same time (month/quarter).   
      
   This same pattern is roughly repeated for all 4 regions.  The   
   "roughly" comes to a bit of a bump with Manchester where the beta is   
   slightly negative for the US-based sightings but VERY negative for   
   the UK-based sightings.  It seems the "whatever" may be even more   
   stretched than we thought.  It maybe only operates in some areas of   
   the UK.  Or maybe people that disappear in London later turn up in the   
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)