NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
Linear Regression In Reverse
From: Peter Fogg
Date: 2005 Jun 5, 08:08 +1000
From: Peter Fogg
Date: 2005 Jun 5, 08:08 +1000
George has written, on the process of LINEAR REGRESSION IN REVERSE: > I just don't understand what Peter is proposing here. My fault, rather > than > his, no doubt. > > Could he provide a bit more detail, please, perhaps with an example? > Couldn't the moment, at which the Sun was theoretically on the East-West > line, be precalculated ... Yes, precalculation is what I am proposing, to give the moment of east or west azimuth (prime vertical). Then a series of observations of the body (it can be any celestial object for which almanac data is available, just as observation of meridian passage can, in theory, be made with any body). From then on it is the application of what seems a favourite hobby-horse of mine, comparing the slope with the sights. Its something I have written about here time and again; once with the title 'Good Data from Bad'. Now there's an approach you can appreciate. The practical advantage here is that it is not necessary to manage to make a timed observation at exactly the precalculated moment. So long as a series of sights encompasses that moment the time can be selected later from the graphed time axis and related, via the slope, to its accompanying altitude on the other axis. Rather than churn out more words I would encourage you to provide your own example, just as I have with Fred. Give it a whirl. I hesitate to say much more; for fear of becoming a bore just banging on about his id?e fixe, but I will say that the moment I understood this slope method was an Eureka moment for me and it has become part of the process of timed sights, just like recording the compass bearing of the body at the time of observation. Good or bad, the only reason I do these things is because they work a treat. How do you like the new name: LINEAR REGRESSION IN REVERSE. Nifty, huh?