Megha Vij
Megha Vij
Aug 31, 2018 · 1 min read

Thank you for this intuitive post!
I wanted to understand that if I wish to interpret the top k coefficients of a regression model, as the likelihood and if I assume a distribution over those k model parameters, can I then apply the Bayesian approach to obtain the posterior probability of a class C in terms of the parameter space as formed by the k parameters?

What I wish to achieve by doing so, is to come up with a posterior distribution of classes in terms of their localised parameter spaces. I would like to seek your advice if this is the correct way to approach such an interpretation.