I'm having trouble understanding how to specify a model for Bayesian inference.

I have a set of measurements, each with known uncertainty, and I would like to infer the mean of the measurements with gaussian model. The mean is known to be within the interval [0,1], so I use an appropriate prior to enforce that. But I'm not sure that I'm handling the measurement uncertainty properly.

The model I have is $$ \begin{align} x_{\textrm{obs},i} &\sim \textrm{Normal}(x_{\textrm{true},i}, \sigma_{\textrm{true},i}) \\ x_{\textrm{true},i} &\sim \textrm{Normal}(\mu, \sigma) \\ \mu &\sim \textrm{Uniform}(0, 1) \\ \sigma &\sim \textrm{Exponential}(0.25) \end{align} $$ where the "obs" variables are the known values and uncertainties for my measurements.

Is this correct?

  • $\begingroup$ Do you have repeated measurements, i.e. multiple $x_{\mathrm{obs},i}$ for each $i$? $\endgroup$
    – Durden
    Commented Jun 17, 2023 at 22:37
  • $\begingroup$ I have known, normally distributed error for each point $\endgroup$
    – qsfzy
    Commented Jun 19, 2023 at 2:46


Your Answer

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

Browse other questions tagged or ask your own question.