Looking for Advice on Using PyMC3's Bayesian Inference in Cognitive Modelling
Hello Everyone,
As I work on this cognitive modelling project, I'm attempting to use PyMC3 to construct Bayesian inference. My aim is to develop a probabilistic model that can aid in the explanation of some cognitive functions, particularly those involved in making decisions in the face of uncertainty. Although I understand the fundamentals of Bayesian inference, I'm having trouble with some of the more intricate parts of putting it into practice with PyMC3 in the community.
These are a few particular problems I'm having:
Model Details: In PyMC3, I'm not sure how to properly define my cognitive model. Even after reading the material, I still have trouble turning my theoretical model into a probabilistic one. Exist any resources or best practices that could be useful in this situation?🤔
Prior Selection: Selecting the right priors for my requirements has proven to be a challenging undertaking. I don't want to impose too much prior information, but I also don't want to be overly ambiguous. How do your cognitive models handle prior selection? 🤔Any advice or general guidelines would be very valued.
Assessment of the Model: What are the best methods for assessing the model's performance once it is constructed? Although I'm aware of techniques like WAIC and LOO, I'm not quite sure how to use them. Exist any guides or illustrations that show how to assess a Bayesian cognitive model correctly?🤔
Any guidance, materials, or real-world examples that could assist me overcome these obstacles would be greatly appreciated. If you have any experience with PyMC3 for cognitive gcp modelling, please share your methodology and any potential difficulties you may have encountered.
Thank you in advance.