crgagne / volatility_paper_elife Goto Github PK

Hi Crgagne,

This is Lili, a PhD student from Dublin City University. I'm really interested in the paper you published on eLief and am so impressed by the modeling analysis you made.

It was said in the left plot of figure 3 that the group mean learning rates are in logit space. I'm confused because I thought if it's in logit, the values should all be in the range 0 to 1. So I am trying to replicate your plots in figure 3. However, it gave me errors when I was trying to import the pickle files of the fitting results (please see the following error codes in the Figures_3-4_Behavioral_Model_Exp1_Results.ipynb file and the screenshot of the error information). I have followed the tutorial created the conda environment and installed the related packages. Could not figure out what was wrong on my side. Could you please help with this? Or at least could you please explain a bit the meaning of the logit space? I'm doing the modeling using Rstan apart from replicating your results with your pymc3 codes, so wanna make sure the Rstan models are correct by comparing the posterior distribution reported in your paper.

model_name = 'model=11_covariate=Bi3itemCDM_date=2020_4_27_samples=2000_seed=3_exp=1.pkl'
with open('../fitting_behavioral_model/model_fits/'+model_name, "rb" ) as buff:
model_output = pickle.load(buff)