When does it make sense to use a bayesian approach, maybe in context to linear regression? To be more concrete: Assume you measure a certain number of devices and you wanna' check the linear relationship between, e.g., voltage and current. Why bayesian and/or why not? How would it be different?



1 Answer 1


The Bayesian approach should be used in the case of:

  1. Strong priors - You have preexisting data and / or domain knowledge that you want to incorporate into the analysis.

  2. Distributional estimates - Instead of point estimates, the result of Bayesian analysis will yield distributions. Those distributions will better quantify the uncertainty of predictions.

In the case of voltage and current for devices, Bayesian linear regression will predict a distribution of current for a given voltage. That is useful because many devices are noisy and possibly the degree of noise changes over voltages.

  • $\begingroup$ Thank you! So, to me this sounds like one should always prefer bayesian over frequency? $\endgroup$
    – Ben
    Dec 8, 2020 at 8:01
  • $\begingroup$ I find the Bayesian framework more useful than frequentist framework most of the time but not always. Frequentist framework is more useful in simple situations and when the software does not support Bayesian computation. $\endgroup$ Dec 8, 2020 at 14:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.