The speed-accuracy trade-off problem is haunting over my mind for a long time these days. I have learned that the most reliable solution is the diffusion model solution. Apart from the complex math behind it, I want to know how many trials are needed at least to fit:
I am so appreciated that the experienced you to give me some hint on it.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2474742/
http://link.springer.com/article/10.3758/BF03194023
http://link.springer.com/article/10.3758/PBR.15.6.1229
you must fit the model to your data not reproduce data predicted by a model. from the second page of yourfirst link. "The experiments also illustrate the strong constraints that ensure the model is empirically testable and potentially falsifiable". Unless this really is psychohistor, what the model gives you is probability of the outcomes. which can change if there is a positive feedback loop or something else that affects the inputs.
Thanks Damora. I guess that you might mean modeling is not a good way to deal with some of the psychology issues. I admit. Nonetheless, do you have any idea about this particular model?
Recently, I found a useful article about this:
https://link.springer.com/article/10.3758%2Fs13428-016-0740-2
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling