Re: [IRCA] Elad FDM S1 + Miscellaneous Mumblings

[Nick Hall-Patch <nhp@xxxxxxxx>]

Actually, Chuck, I worked quite a lot with Bob when I was doing the 
MW signal strength recorder article for QEX, yet he never mentioned 
regression analysis as a possible way of making sense of my signal 
strength measurements vs. any number of solar/geomagnetic 
parameters.   There is no argument that he would have been extremely 
familiar with it, hence my later assumption that he didn't think 
there was much to be gained by using it on this particular data 
set.   He certainly pushed me in other directions such as spherical 
geometry and ionospheric chemistry that made my not particularly 
mathematical head hurt from time to time. Although I believe he tried 
to make a point of not larding his propagation articles with 
mathematics, I don't think he had any qualms about stretching the 
brains of interested individuals, so, as you describe it, "standard 
stuff" such as regression analysis would have been suggested if he 
thought there was merit in it.

Have you yourself ever tried to derive relationships between any of 
the parameters Michael mentioned and signal strength levels or times 
of peak signals?   It would likely be easiest to start with local 
sunrise times vs. time of peak signal strength during sunrise / dawn 
enhancement, though I suspect most of us would be happy to continue 
with our rules of thumb about that, but even then, there are further 
variables.  DU's can deliver post-sunrise reception on the coast in 
summer, but generally not inland for example, and that doesn't seem 
entirely due to lower signal levels inland.

And, thanks for mentioning the linear regression capabilities of 
Excel.  The few times I've attempted to use it, a simple slope and 
offset gave as good a correlation as anything higher level, but it's 
good to be reminded that it is there.

best wishes,

Nick


At 08:05 31-12-13, you wrote:
> Nick:
>
> Paragraph #1: Linear regression is built into Excel so the world can enjoy it.
>
> Regression analysis to understand the degree of correlation between 
> variables is very standard stuff. If Bob didn't mention it, it's 
> because he didn't mention basic techniques of statistical analysis.
>
>
> Chuck
>
> >


_______________________________________________
IRCA mailing list
IRCA@xxxxxxxxxxxxxxxx
http://montreal.kotalampi.com/mailman/listinfo/irca

Opinions expressed in messages on this mailing list are those of the original contributors and do not necessarily reflect the opinion of the IRCA, its editors, publishing staff, or officers

For more information: http://www.ircaonline.org

To Post a message: irca@xxxxxxxxxxxxxxxx