Are you using adjusted r^2 here?
Alan
>
> With an all London axial map converted to segments:
>
> Total Angle to Total Angular Depth
> R2 (180 degree turn) R^2 = 0.999
> R3 R^2 = 0.998
> R4 R^2 = 0.996
> R5 R^2 = 0.993
>
> There is the gradual curving off as you have shown for the topological
> case.
>
> However, intriguingly, it also correlates to a strong degree with the
> node count:
>
> Node Count to Total Angular Depth
> R2 R^2 = 0.997
> R3 R^2 = 0.996
> R4 R^2 = 0.994
> R5 R^2 = 0.990
>
> Alasdair
>
> sheep dalton wrote:
> >> It should probably be added that using angular segmental analysis, the
> >> total angle to radius n does correlate with the total angular depth to
> >> radius n in a similar manner to the topological measure.
> >
> >
> > Yes but how closely ? the current RRA equations never get less than r
> > squared of 0.99* for the axial radius case ( topological). A
> > correlation of 0.8 would be very strong but still not strong enough.
> >
> > sheep
> >
> > ps. Its nice to see the kinds of general academic chit chat this mail
> > base was intended to convey.
> >
> >
> >
> > for our less technical readers a r squared of 1.0 is a perfect
> > correlation say of your hight in meters with your hight in feet. A
> > correlation of 0.0 means totally unrelated for example the numbers in
> > you phone book against the production of rice in Indonesia.
>
> --
> Alasdair Turner
> Course Director
> MSc Adaptive Architecture and Computation
> Bartlett School of Graduate Studies
> UCL Gower Street LONDON WC1E 6BT
>
> http://www.aac.bartlett.ucl.ac.uk/
|