There are various way to deal with data series that are "sparse".  The   
   s/w has run through most of them and the results are broadly   
   consistent.   
      
   But I'll present here a typical transformation of the data. We convert   
   from events that happen every few days -- e.g. a list of months with a   
   low count of events in that month -- to a month by month estimate of   
   the "return time". I.e. the number of months between 1 event and the   
   next at this point in time.   
      
   With a month-by-month estimate of return times we can use all the   
   machinery of handling real numbers instead of small integers --   
   e.g. interpolating to get estimates of missing months.   
      
   Running these against the now 28k time series in the database gives us   
   a list with the top 20 items as follows:   
      
   Suspect variable	Trans		R2   
   cosmic-ERVN		0o1		0.87358840   
   cosmic-MCMD		1o1-y		0.85402629   
   gavcosmic-ERVN		0o2		0.84920654   
   cosmic-FSMT		3o1-y		0.84565447   
   cosmic-FSMT		0o1-y		0.84231633   
   gavcosmic-MWSN		0o1		0.83703157   
   cosmic-THUL		3o1		0.82952817   
   gavcosmic-SNTG		3o1-y		0.82803157   
   cosmic-INVK		3o1		0.82683286   
   cosmic-INVK		3o1-y		0.82575344   
   cosmic-NAIN		0o1-y		0.82079697   
   cosmic-ERVN		0o2		0.81983241   
   gavcosmic-ERVN		0o1		0.80868246   
   gavcosmic-THUL		0o1-y		0.80534677   
   gavcosmic-NAIN		1o1-y		0.78479206   
   gavcosmic-NAIN		0o1-y		0.78267140   
   cosmic-THUL		12o1-y		0.77392370   
   gav5904179		1o1		0.76797919   
   gavcosmic-MWSN		0o2		0.76743680   
   gav5904179		1o1-y		0.76681924   
      
   The first column gives the code name of the data series being   
   correlated against the return times for monthly jelly UFO sightings.   
   The 2nd column is a code for what transformations the AI found best   
   tuned the match to the best possible it could find.  And the 3rd   
   column gives the "explanation power" of the series named in col 1.   
      
   "Cosmic-ERVN" is the monthly output from a cosmic ray telescope in the   
   mountains of Armenia. It has shown up in quite a few UFO-related   
   studies I've posted in the past few months. If we look in detail at   
   the actual model, it shows more cosmic rays are associated with a   
   larger return time -- i.e. greater time between sightings.  It seems   
   the "sky jellyfish" don't like flying around when lots of cosmic   
   radiation is coming down from the sky.   
      
   We note from table almost all the top 20 are related to cosmic rays.   
   Jellyfish *really* don't like radiation.  The major difference is which   
   station in the cosmic telescope community did the measurements. While   
   stations are scattered all around the world it seems the majority of   
   the ones in the list, above, are around the poles.   
      
   The only other items mentioned in the table is the mystery number   
   5904179. And that is a robot in the southern ocean that moves around,   
   up and down, to measure ocean water temperature, salinity and   
   chlorophyll content.  Somehow the water temps down there in the   
   Southern Ocean explain about 76% of the month-to-month variation in   
   estimated return times for jellyfish sightings.   
      
   Given there are 10s of 1000s of data series and the above list is   
   only 20 of them, we have to look further to see more interesting   
   patterns.  E.g. we can ask "are any planets implicated"?   
      
   And we get a list like this:   
      
   Planet		Variable	Trans		R2   
   gavjupiter	RA		3o1-y		0.41710465   
   gavmars		rg		12o1-y		0.20605815   
   gavmercury	RA		0o1		0.08690225   
   gavpluto	elong		1o1		0.06147527   
   gavsaturn	v		12o1-y		0.17084634   
   gavsol		r		1o1		0.03720755   
   gavuranus	r		3o1-y		0.13934915   
   gavvenus	RA		0o1		0.05865361   
   gavvenus	RA		0o1		0.05865361   
      
   I.e. the Right Ascension of Jupiter (with the appropriate   
   transformation listed -- related to removing very distant outliers,   
   lagging the series by 3 months, and using a log transform on the   
   return times) explains around 42% of the month-to-month variation in   
   return times. I.e. when Jupiter moves to a "good" celestial longitude   
   about 3 months later jelly sightings seem to react.   
      
   And we can also ask what things other than cosmic rays and planetary   
   positions related to jellyfish sightings:   
      
   Suspect			Trans		R2   
   gavocean-70dep1		0o1		0.74777184   
   gavocean-60dep10	3o1		0.67552692   
   gavocean50dep500	1o1		0.66729375   
   gavocean-60dep500	3o1		0.66368744   
   gavocean-60dep1000	3o1		0.65024259   
   ocean-60dep100		3o1		0.63115045   
   gavocean-70dep10	3o1		0.51762137   
   gavocean50dep100	3o1		0.49251600   
   gavocean-10dep0		0o1-y		0.47532055   
   gavocean-60dep1		0o1-y		0.42208783   
   ocean50dep1000		1o2		0.38352995   
   ocean-20dep500		12o1-y		0.37085697   
   gavn2o_mhd		0o1-y		0.34706927   
   gavqjapan		12o1		0.32914547   
   gavwindband-60		12o1-y		0.32569189   
   gavocean0dep10		3o1-y		0.32521161   
   gavocean-40dep1000	12o1-y		0.32134049   
   gavpresseg-10		12o1-y		0.30766505   
   gavpresband-60		12o1-y		0.29243382   
   gavqbrazil		1o1-y		0.28586552   
   gavqseg-50		1o1-y		0.28586552   
   ocean-10dep1		1o1-y		0.27751417   
   qseg30			3o1		0.27439373   
   ocean-30dep500		12o1		0.25191674   
   gavn2o_thd		3o1-y		0.24386299   
   gavn2o_kum		0o1-y		0.22749707   
   gavocean-10dep500	12o1		0.21927441   
   gavarc-70		1o1-y		0.21506766   
   gavn2o_cgo		0o1-y		0.21465639   
   gavcemband-6		12o1		0.21056614   
   gavphytoindexS		12o1-y		0.21054219   
   gavocean30dep1000	3o1-y		0.20943744   
   gavcemband-5		12o1		0.20939120   
   gavocean0dep500		12o1		0.20926377   
   gavocean-40dep500	12o1		0.20216449   
   gavworld60		0o1-y		0.19634182   
   gavarc-60		1o1-y		0.19407190   
   gavocean-20dep100	12o1-y		0.19379212   
   gavn2o_sum		0o1-y		0.19163500   
   gavn2o_psa		3o1-y		0.18781128   
   gavocean-70dep100	12o1-y		0.18395385   
   gavmars-FV		12o1-y		0.18313543   
   gavqaustralia		12o1		0.18152462   
   gavworld50		1o1-y		0.17998850   
   ocean-70dep500		0o1		0.17120217   
   qbandmag-30		1o1-y		0.16182924   
   gavant20		0o1-y		0.15758243   
   gavant30		0o1-y		0.15697096   
   gavn2o_nwr		1o1-y		0.15605519   
   gavwindseg-90		12o1-y		0.15486991   
   gavarc20		12o1-y		0.15388346   
   nh			0o1-y		0.15324586   
   gavaravgEquat		3o1-y		0.15039821   
   gavn2o_NH		1o1-y		0.14930516   
   gavarc50		12o1-y		0.14845662   
   gavwindseg-30		0o1		0.14679843   
   gavocean50dep10		0o2		0.14678671   
   gavaravgSHocean		0o1		0.14183093   
   lat-10			0o1-y		0.14130550   
   gavocean-40dep1		12o1-y		0.14032628   
   gavcemband-7		12o1-y		0.13092820   
   gavocean30dep500	3o1-y		0.13091940   
   ocean-70dep1000		0o1		0.13078407   
   gavaravgSH		0o1		0.12878416   
   gavworld70		12o1-y		0.12790891   
   gavqrussia		3o1		0.12789781   
   gavocean0dep100		3o2		0.12673678   
   mars-elong		0o1		0.12550588   
   gavqband30		12o1		0.12423087   
   gavant40		0o1-y		0.12329068   
   gavqindonesia		3o1		0.12170414   
   gavlat-60		12o1		0.12128601   
   gavqnatlridge		0o1-y		0.12012003   
   windseg-20		1o1-y		0.11982142   
   gavocean40dep1000	3o1-y		0.11978953   
   gavocean-30dep10	3o1-y		0.11845415   
   gavpresband10		3o1-y		0.11745526   
   gavindia		3o1-y		0.11694643   
   gavqwindantridge	12o1-y		0.11647169   
   gavqvenezuela		12o1-y		0.11343549   
   uah_globe6NH		0o1-y		0.11275515   
   gavpresband20		3o1-y		0.11172239   
   gavocean0dep1		0o1		0.11136655   
   gavuah_mtNHOcean	1o1-y		0.11039428   
   ocean0dep0		12o1-y		0.10977220   
   gavwindseg-130		1o1-y		0.10846157   
   gavqoman		1o1		0.10813321   
   gavsunspots		12o1		0.10807520   
   gavworld80		0o1-y		0.10688617   
   lat-20			1o1		0.10613056   
   gavvolcano_dates	12o2-y		0.10532982   
   gavant-30		12o1		0.10511143   
   gavarc130		12o1-y		0.10446238   
   gavocean10dep1		12o1-y		0.10365002   
   qbandmag40		12o1-y		0.10312364   
   gavpresseg-30		0o1		0.10241679   
   gavwindseg30		12o1-y		0.10233889   
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)