starting point parameter in pstRT_rlddm6()?

[xinx...@126.com]

 

Dear list,

Thank you very much for reading this message.

Does the pstRT_rlddm6() function provide the parameter estimate for the starting point? If not, why?

Here is what I have got with the example code:

````

output <- pstRT_rlddm6(data = "example", niter = 2000, nwarmup = 1000, nchain = 4, ncore = 4)

output$fit

Inference for Stan model: anon_model.
4 chains, each with iter=2000; warmup=1000; thin=1; 
post-warmup draws per chain=1000, total post-warmup draws=4000.
                mean se_mean   sd     2.5%      25%      50%      75%    97.5% n_eff Rhat
mu_a            1.90    0.00 0.09     1.73     1.84     1.90     1.96     2.08  1342    1
mu_tau          0.36    0.00 0.01     0.34     0.35     0.36     0.36     0.37  3800    1
mu_v            4.55    0.01 0.44     3.72     4.26     4.56     4.84     5.43  3733    1
mu_alpha        0.07    0.00 0.02     0.03     0.05     0.06     0.08     0.12  1123    1
sigma[1]        0.14    0.00 0.04     0.08     0.11     0.13     0.16     0.24  1567    1
sigma[2]        0.08    0.00 0.06     0.00     0.03     0.07     0.12     0.23  2232    1
sigma[3]        0.61    0.00 0.18     0.22     0.49     0.61     0.73     0.94  1924    1
sigma[4]        0.54    0.00 0.09     0.38     0.47     0.53     0.59     0.74  2498    1

......

````

Thanks.

YA

[Jeung-Hyun Lee]

Hi YA, 

Sorry, I’m not entirely sure what you mean by the "starting point," but typically, the initial value of a parameter estimate is set randomly within the specified range for each parameter. This approach helps minimize the potential bias from the initial values on the parameter estimation. I hope this clarifies things, and please feel free to reach out if anything is unclear!

Best, 

JH 

[ya]

hi JH,

Thanks for the response.

By starting point, I meant the initial value z, in DDM language:)

YA

---- Replied Message ----

From	Jeung-Hyun Lee<jeungh...@gmail.com>
Date	08/11/2025 12:59
To	hbayesdm-users <hbayesd...@googlegroups.com>
Subject	Re: starting point parameter in pstRT_rlddm6()?

--
You received this message because you are subscribed to the Google Groups "hbayesdm-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hbayesdm-user...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/hbayesdm-users/de5c0626-73be-40f8-b9d8-369bffb54256n%40googlegroups.com.

[Simon]

Hi YA,

I guess you’re talking about the starting parameter z (often also described as the bias parameter beta) of the Wiener distribution. This is fixed to 0.5 in pstRT_rlddm6() , as is usually the case when using the Wiener distribution as the choice rule in reinforcement learning. This is because it is not assumed that there was an a-priori bias towards one of the choice options in two alternative forced choice learning paradigms. Instead, it is assumed that the feedback for taking choices (as provided to pstRT_rlddm6()via the data) is creating a preference for a choice option which is then captured by the drift rate v (group level hyper parameter mu_v in your output), that varies from trial to trial.

So the answer to your question is no, the starting point parameter is not reported because it is not estimated by the model. Instead it is fixed to 0.5.

Does that answer your question?

Best, Simon

[ya]

hi Simon,

Thank you very much for the clarification. Now I am clear.

Best regards,

YA

---- Replied Message ----

From	'Simon' via hbayesdm-users<hbayesd...@googlegroups.com>
Date	08/11/2025 15:46

To	hbayesdm-users<hbayesd...@googlegroups.com>
Subject	Re: starting point parameter in pstRT_rlddm6()?

To view this discussion visit https://groups.google.com/d/msgid/hbayesdm-users/F9D0CD8D-F5C1-493F-B0AB-9DDDBF5D6BA2%40gmx.de.