tag:blogger.com,1999:blog-4297242917419089261.post3333391857757183311..comments2024-03-09T01:05:10.754-08:00Comments on Babies Learning Language: Explorations in hierarchical drift diffusion modelingMichael Frankhttp://www.blogger.com/profile/00681533046507717821noreply@blogger.comBlogger7125tag:blogger.com,1999:blog-4297242917419089261.post-75133966563011399832018-03-23T09:28:03.975-07:002018-03-23T09:28:03.975-07:00I didn't know about this! Thank you. Very help...I didn't know about this! Thank you. Very helpful. Michael Frankhttps://www.blogger.com/profile/00681533046507717821noreply@blogger.comtag:blogger.com,1999:blog-4297242917419089261.post-29366770640978650342018-03-23T00:57:08.141-07:002018-03-23T00:57:08.141-07:00You probably know these already, but here are a co...You probably know these already, but here are a couple of newish implementation of the model in stan, which work pretty fine (tho' or maybe because btw trial variability is not included):<br />- http://singmann.org/wiener-model-analysis-with-brms-part-i/<br />- https://github.com/cran/hBayesDM/blob/master/inst/stan/choiceRT_ddm.stanRiccardohttps://www.blogger.com/profile/13675131893260129288noreply@blogger.comtag:blogger.com,1999:blog-4297242917419089261.post-15826971688730847342017-07-24T08:57:57.774-07:002017-07-24T08:57:57.774-07:00This comment has been removed by the author.Samir Huseynovhttps://www.blogger.com/profile/03712893304495430563noreply@blogger.comtag:blogger.com,1999:blog-4297242917419089261.post-36007455037003921002016-02-18T21:14:03.056-08:002016-02-18T21:14:03.056-08:00Thanks very much, Joachim. In retrospect I probabl...Thanks very much, Joachim. In retrospect I probably should have started with JAGS - but having switched to Stan I am now more used to its syntax etc. <br /><br />Thanks also for all the work you have done on these software implementations, I know the research community appreciates it!<br /><br />Agreed about the trial-level variability issue, also. I suspect this will be a moot point in our dataset anyway, as I have just read Ratcliff and Childer's HDDM simulations that suggest this variability is undetectable with our number of trials (http://star.psy.ohio-state.edu/coglab/People/roger/pdf/lownfinaldec14.pdf).<br />Michael Frankhttps://www.blogger.com/profile/00681533046507717821noreply@blogger.comtag:blogger.com,1999:blog-4297242917419089261.post-78199343198933966492016-02-18T21:07:10.946-08:002016-02-18T21:07:10.946-08:00Nice write-up! I just wanted to add that the Stan...Nice write-up! I just wanted to add that the Stan implementation isn't my favorite (as I pointed out in that Google group, the odd omission of the binary outcome as a parameter is a problem for several reasons). You can probably speed it up by splitting the data set into a "corrects" and "errors" vector before feeding it to Stan, and then dispense with the if-block.<br /><br />That said, until the Stan implementation gets looked at properly, the JAGS implementation is the most stable/useful around.<br /><br /><br />There's a lot to be said about the trial-to-trial variability parameters, but I tend to avoid them as much as I can. They're not only numerically tricky, they're hard to identify, and if you implement them as we did in one example (with unique nodes for each trial) you risk building a near-unfalsifiable model (which can still be useful for measurement, mind you...).Anonymoushttps://www.blogger.com/profile/00457442016473274970noreply@blogger.comtag:blogger.com,1999:blog-4297242917419089261.post-28106795010135707952016-02-18T14:39:28.836-08:002016-02-18T14:39:28.836-08:00Yes, definitely could be sampler, or sampler imple...Yes, definitely could be sampler, or sampler implementation, or the model structure. Very tough to tell because I wasn't able to match model forms exactly. Michael Frankhttps://www.blogger.com/profile/00681533046507717821noreply@blogger.comtag:blogger.com,1999:blog-4297242917419089261.post-19171393823289677492016-02-18T13:34:23.300-08:002016-02-18T13:34:23.300-08:00Awesome! Thanks for sharing these explorations. Co...Awesome! Thanks for sharing these explorations. Could the Stan vs. PyMC speed difference be due to NUTS vs Metropolis?Anonymousnoreply@blogger.com