Hello there, im new to the probabilistic programming, but i find it very interesting. I’m well-versed in classic machine learning approaches, but I’m looking to introduce an additional dimension of uncertainty and explainability in my project. I want to estimate soil nutrient deficiency in my fields by combining satellite imagery (e.g., NDVI index), electrode data, images of plants, spectral data, and soil properties data using Bayesian inference. How can I effectively implement Bayesian methods to incorporate uncertainty and enhance the explainability of my nutrient deficiency estimates? I need some directions that will help me to successfully do this
1 Like
Hi Remo!
Do you have a specific model or reference paper in mind to implement? PyMC is sort of a blank canvas. If you can give some details about what you want to do, we can try to give some suggestions about how to do it.
In general, if you’re new to attacking scientific problems from a Bayesian perspective, I suggest looking at the Statistical Rethinking, which is a book (can’t link that) with video lectures and accompanying PyMC code (book code / video code).
Otherwise the PyMC examples library is a great place to go to get some ideas. I recommend studying the Radon case study as an introduction to hierarchical modeling.
.
1 Like