NavList:

   

Reply

Perhaps I'm confused, but why wouldn't it be reasonable to derive the
regression coefficients from the data?  What "approximate environment"
data would be used to "precompute" the coefficients?

On Sat, October 13, 2018 10:30, Ron Jones wrote:
> Dear Antoine
> You wrote
> __________________________________________________________________________
> _________________________________________________________________________
> ________________________________
> "In practice and from your own experience how does your 2nd order
> regression perform since you seem to be using the (x2) 2nd order
> coefficient *directly derived from the data set* and not precomputed from
> the Observer's (approximate) environment ? (*RE:
> Linear-regression-other-tools-Couëtte-oct-2018-g43055)*
> In other words, does your 2nd order regression bring a definite practical
> advantage over the 1st order [linear] one ?"
> _________________________________________________________________________
> _________________________________________________________________________
> __________________________________
> For a string of sights taken near meridian transit (+/- 10 minutes) the
> second order regression may provide a better result.  For a string of
> sights taken over a short time before & after meridian transit use the
> Meridian Transit worksheet which uses second order regression; and
> calculates the time and sextant altitude associated with meridian transit
> (see attached PDF)
> Ron ..... 48 N  123 W
>
>
> [plain text auto-generated]
> ----------------------------------------------------------------
> NavList message boards and member settings: http://fer3.com/NavList
> Members may optionally receive posts by email.
> To cancel email delivery, send a message to NoMail[at]fer3.com
> ----------------------------------------------------------------
> ----------------------------------------------------------------
>
>
>
> Attached File: http://fer3.com/arc/imgx/Meridian-Transit-Worksheet.pdf
>
>
>
> :
> http://fer3.com/arc/m2.aspx/Linear-regression-other-tools-RonJones-oct-201
> 8-g43060

   

Reply