I've heard a couple of definitions of a Bayesian plot, but I am not 100% sure what it is and whether it is possible to plot this in an R.

An example of the plot is here: http://www.springerimages.com/Images/MedicineAndPublicHealth/1-10.1007_s11307-008-0154-3-1

In this particular example there are four different tests, the prior probability is evaluated based on the proven probability of cancer recurrence based on 7 patients. This is compared against varying prevalences of tumour (As far as I can understand this is the the x-axis reflects these changing prevalences). The y-axis is the posterior probability. What I don't fully understand is what the likelihood is and how they get it from the different methods to actually calcualte the posterior and show there are differences between these methods.

I am not necessarily interested in understanding this particular example. I just want to understand how I can construct and interpret this plot on my own data. I understand that my prior should be my disease prevalence, but what I don't understand is how the likelihood should be represented in this plot (is it the x-axis?). From my own data I can evaluate the posterior because I know the prior and the likelihood... I just don't know how to put them into this plot.

Has anyone made such a plot for their own work that can help me to understand the intuition?



Your Answer

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