Rachlin's hyperboloid function is df = 1/(1+k(D^s)). However, as pointed out by Vincent & Stewart (2018) it is incorrectly parameterised. So we will implement the modified Rachlin discount function, as proposed by Vincent & Stewart (2018). This is df = 1/(1+(kD)^s).
[x] create the new ModifiedRachlin class
[x] add it to tests
[x] add it to the PsychoPy demo
[x] add it to the notebook on multiple model fitting
[x] update README.md
References
Rachlin, H. (2006, May). Notes on Discounting. Journal of the Experimental Analysis of
Behavior, 85(3), 425–435.
Vincent, B. T., & Stewart, N. (2018, October 16). The case of muddled units in temporal discounting. https://doi.org/10.31234/osf.io/29sgd
Rachlin's hyperboloid function is
df = 1/(1+k(D^s)). However, as pointed out by Vincent & Stewart (2018) it is incorrectly parameterised. So we will implement the modified Rachlin discount function, as proposed by Vincent & Stewart (2018). This isdf = 1/(1+(kD)^s).ModifiedRachlinclassReferences
Rachlin, H. (2006, May). Notes on Discounting. Journal of the Experimental Analysis of Behavior, 85(3), 425–435. Vincent, B. T., & Stewart, N. (2018, October 16). The case of muddled units in temporal discounting. https://doi.org/10.31234/osf.io/29sgd