Multiple
Regressors: Near Collinearity
¥Red curve = signal model for class
#1
¥Green curve = signal model for
#2
¥Blue curve =
b1×#1+(1Ðb1)×#2
where b1 varies
randomly from 0.0 to 1.0 in
animation
¥Gray curve =
0.66×#1+0.33×#2
= simulated data with
no noise
¥ Lots of different combinations of #1 and #2 are decent fits to gray curve
Stimuli are too close in time to
distinguish
response #1 from #2,
considering noise
Red & Green stimuli average 2 s apart