NavList:
A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
Re: Lat/Lon by "Noon Sun" & The Noon Fix PROVE IT
From: James N Wilson
Date: 2009 Apr 29, 08:56 -0700
From: James N Wilson
Date: 2009 Apr 29, 08:56 -0700
Brad: I'm sure that least squares is better, but I used regression analysis because that was what I had. I'm a complete novice at this mathematical curve fitting. I was grasping at straws to avoid asking for more data points. I've looked at a good many student sight plots, and they are almost uniformly on a straight line. But they are taken with a sharp horizon. I suspect that a hazy horizon would yield the kind of data that George generated. Being curious, I immediately tried this new toy on my ancient data example. It placed the ascending line in exactly the same place that I had eyeballed it, and had only a slight difference in the descending line. The net result was a slight improvement in accuracy. In eyeballing sights, I tend to throw out ones quite a ways from the trend, whereas any curve fitting algorithm will use them all. In looking at the final results, sights taken in runs centered around LAN would be more utile than ones taken around highest altitude. That would eliminate the extrapolation. But, as George points out, the observer doesn't know when LAN is yet. More sights could be taken, so that the latter ones would avoid extending the data lines outside of the range measured. I think that I still have a lot to learn. Jim Wilson ____________________________________________________________ Save $10 on Flowers and Gifts! Shop now at www.ftd.com/16714 --~--~---------~--~----~------------~-------~--~----~ Navigation List archive: www.fer3.com/arc To post, email NavList@fer3.com To , email NavList-@fer3.com -~----------~----~----~----~------~----~------~--~---