digicosmos86
commented
2 months ago Hi @Hellobamboobamboo
As a convention we only accept responses coded as -1 and 1 for the binary case, to differentiate with cases with more than two responses, where we accept 0, 1, 2, and so on. Once the coding is changed I think you'll be able to proceed to plotting
I am also going to move this to discussions since this is not a bug
Describe the bug A clear and concise description of what the bug is.
HSSM version 0.2.3
I was running my model like this: basic_model = hssm.HSSM( 7 model="ddm", # Specify the model type as drift diffusion model (DDM) 8 data=data, 9 include=[ 10 { 11 "name": "v", 12 "formula": "v ~ current + diff + condition + status + gender", 13 "prior": { 14 "Intercept": {"name": "Normal", "mu": 0.0, "sigma": 1.0}, 15 "current": {"name": "Normal", "mu": 0.0, "sigma": 1.0}, 16 "diff": {"name": "Normal", "mu": 0.0, "sigma": 1.0}, 17 "condition": {"name": "Normal", "mu": 0.0, "sigma": 1.0}, 18 "status": {"name": "Normal", "mu": 0.0, "sigma": 1.0}, 19 "gender": {"name": "Normal", "mu": 0.0, "sigma": 1.0}, 20 }, 21 "link": "identity", 22 }, 23 { 24 "name": "a", 25 "formula": "a ~ current + diff + condition + status + gender", 26 "prior": { 27 "Intercept": {"name": "Normal", "mu": 0.5, "sigma": 1.0}, 28 "current": {"name": "Normal", "mu": 0.0, "sigma": 1.0}, 29 "diff": {"name": "Normal", "mu": 0.0, "sigma": 1.0}, 30 "condition": {"name": "Normal", "mu": 0.0, "sigma": 1.0}, 31 "status": {"name": "Normal", "mu": 0.0, "sigma": 1.0}, 32 "gender": {"name": "Normal", "mu": 0.0, "sigma": 1.0}, 33 }, 34 "link": "identity", 35 } 36 ], 37 )
then I get a error like this:
File ~\AppData\Local\anaconda3\envs\hssm_env\Lib\site-packages\hssm\hssm.py:383, in HSSM.init(self, data, model, choices, include, model_config, loglik, loglik_kind, p_outlier, lapse, hierarchical, link_settings, prior_settings, extra_namespace, missing_data, deadline, loglik_missing_data, process_initvals, **kwargs) 380 self.p_outlier = self.params.get("p_outlier") 381 self.lapse = lapse if self.has_lapse else None --> 383 self._post_check_data_sanity() 385 self.model_distribution = self._make_model_distribution() 387 self.family = make_family( 388 self.model_distribution, 389 self.list_params, 390 self.link, 391 self._parent, 392 )
File ~\AppData\Local\anaconda3\envs\hssm_env\Lib\site-packages\hssm\hssm.py:1683, in HSSM._post_check_data_sanity(self) 1679 if np.any(~np.isin(unique_responses, self.choices)): 1680 invalid_responses = sorted( 1681 unique_responses[~np.isin(unique_responses, self.choices)] 1682 ) -> 1683 raise ValueError( 1684 f"Invalid responses found in your dataset: {invalid_responses}" 1685 ) 1687 if len(unique_responses) != self.n_choices: 1688 missing_responses = sorted(np.setdiff1d(self.choices, unique_responses))
ValueError: Invalid responses found in your dataset: [0]
what should my response variable's value be, it's usually just 0 and 1? Also, do I still need to flip the rt time to negative manually like in hddm to plotting? But HSSM doesn't take negative as valid this time. What values can it take?