# How to create Bayesian data fusion in python with pymc3?

**URL:** <https://discourse.pymc.io/t/how-to-create-bayesian-data-fusion-in-python-with-pymc3/10558>\
**Category:** v3\
**Created:** [October 7, 2022, 7:20pm UTC](https://discourse.pymc.io/t/how-to-create-bayesian-data-fusion-in-python-with-pymc3/10558 "2022-10-07T19:20:14Z")\
**Posts on this page:** 1\
**Showing post:** 7

<div class="post-metadata">

**Author:** ![Mojo\_ild](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.pymc.io/mojo_ild/32/4703_2.png) [@Mojo\_ild](https://discourse.pymc.io/u/Mojo_ild)\
**Post date:** [October 10, 2022, 12:56pm UTC](https://discourse.pymc.io/t/how-to-create-bayesian-data-fusion-in-python-with-pymc3/10558/7 "2022-10-10T12:56:35Z")

</div>

I tried to create the Bayesian inference based on the information that I found in the following links:

 ![image](https://canada1.discourse-cdn.com/flex036/uploads/pymc3/original/2X/3/3b53a37aa979b2b2f369002f7469e0fba1188985.png)

- [From this link: [Hyperbolic Secant Distribution](https://www.randomservices.org/random/special/HyperbolicSecant.html)]
- [How to choose a Bayesian prior](https://www.healthcare-economist.com/2019/04/04/how-to-choose-a-bayesian-prior/)
- [Prior Choice Recommendations](https://github.com/stan-dev/stan/wiki/Prior-Choice-Recommendations)
- Custom distribution can be found here: [Creating a Custom Hypersecant distribution](https://discourse.pymc.io/t/creating-a-custom-hypersecant-distribution/10255)

I think I created the mu as Generic weakly informative prior and same for sigma (hopefully 😅), but when it comes to creating the fusion part, I keep on coming back to the priors as I include the observed data as distribution data from sensors. Am I right to think and assume that difference in the reuslts is due to my setting up the model for Bayesian fusion? Like the parametrization like `target_accept`, `draws`, `tune` and the priors.

---

_[View the full topic](https://discourse.pymc.io/t/how-to-create-bayesian-data-fusion-in-python-with-pymc3/10558)._