Topic 6 Flashcards
What is regression through the origin?
Where we assume beta1 = 0, so our model becomes:
How does our estimate formulas change for regression through the origin?
How does our coefficient of determination change in regression through the origin?
r^2 becomes junk, instead we use raw r^2, which isn’t comparable
When we scale Yi by w1 to Y*i, and Xi by w2 to X*i, how does things change our estimate formulas and other things?
What is this model?
log-log or log-linear (not log-lin)
In a log-log model, what does beta-2 represent?
The percentage change in Y that is expected from a percentage change in X
What is this model?
log-lin (not log-linear)
What does beta2 represent in the log-lin model?
The percentage change in Y that is expected from an absolute change in X
What model is this?
lin-log
What does beta2 represent in the lin-log model?
The absolute change in Y expected from a percentage change in X
Show the reciprocal model?
In the reciprocal model, where does Yi converge to?
The horizontal line Y=Beta1
In the reciprical model, what determines the slope of the SPR?
Beta2^. Positive Beta2^ means downward sloping, Negative Beta2& means upward sloping
What model is this?
An intrinsically non-linear model, which we cannot work with and which must be solved by algorythms on computers. And also, it smells.
