NavList:

   

Reply

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
-~----------~----~----~----~------~----~------~--~---

   

Reply