# uninformative prior for something dependent on states?

Say I have a model which predicts something say abc over time

I define 3 states, S1, S2, S3. I define transition transitional probs.

I have another variable say, xyz which has value of 0 to 5.

xyz will be part of data going into the model. By my own definition, if xyz is between:

• 0-1, suggests abc is in state S1
• 1-3, suggests abc is in state S2
• 4-5, suggests abc is in state S3

xyz[i] ~ dbin(prob.xyz[state[i]],5)

What would be the uninformative prior for prob.xyz[1]?