Is there a way to fix the parameter in the LAN extension model?

[g mq]

Dear HDDM developers and experts,

Sometimes the model is too complex and lacks parameter identification power. Fixing parameters might be an option to increase the parameter identification power. For example, there are some critics that Weibull CDF might be too flexible, fixing the beta parameter in Weibull CDF might increase the parameter recovery, I am wondering if is there a way to fix the parameter in LAN extension model.

 I have read the Network Inspectors in the HDDM doc page, it seems users can access the MLP model and then fix some parameters, using the custom mcmc sampler or parameter optimization algorithm to estimate the parameter. But that would be quite a lot of work. I am just curious is there a way to fix parameters and still use the HDDM original sampler? Also, even if there is no such function now, it would be very easy to add it, right? I think it would be amazing if the next version hddm has this function. Thanks.

Best,

Mingqian Guo

[Michael J Frank]

yes it is one of the features we have been planning to add for some time to let users choose their own priors - if you want to fix a parameter you would just choose the prior to be fixed at that level with 0 variance. But one thing you can do already (which is probably better and recovery should improve) is instead of fixing the parameter just set it to be estimated only across the whole group of subjects, which will fix it for each individual to the group mean. (eg . for weibull do "group_only_nodes = ['beta']" ).

--
You received this message because you are subscribed to the Google Groups "hddm-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hddm-users+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hddm-users/65824ea1-8920-4f60-9421-ea3e6cb874ban%40googlegroups.com.

[g mq]

Thanks for your answer!