contributions, aivi, which is the effect of each reinforcer as remembered.
This derivation gives the predicted effects of varying t and s on the
parameters of the hyperbolic model. The second important result of the
derivation is that our use of the additive noise model yields a value of a
sample that fits a hyperbolic model.
By Equation 13, which states that δi = δ0 + itαsβ , and Equation 2, 10,
Equation 21, which is derived from Equation 20, shows that the decremen-tation function is hyperbolic in form: