I am going through this paper. I am unable to understand the definition of a confidence distribution.
A statistic C is an exact confidence distribution for a real parameter ρ if:
- ρ → C(ρ; y) is a cumulative distribution function for all y in the sample space of the data Y.
- C(ρ; Y ) ∼ U(0, 1)
[Wikipedia] Θ is the parameter space of the unknown parameter of interest θ, and χ is the sample space corresponding to data Xn={X1, ..., Xn}. A function Hn(•) = Hn(Xn, •) on χ × Θ → [0, 1] is called a confidence distribution (CD) for a parameter θ, if it follows two requirements:
(R1) For each given Xn ∈ χ, Hn(•) = Hn(Xn, •) is a continuous cumulative distribution function on Θ;
(R2) At the true parameter value θ = θ0, Hn(θ0) ≡ Hn(Xn, θ0), as a function of the sample Xn, follows the uniform distribution U[0, 1].
I have been studying from this textbook also.
The things that confuse me are that it's not a probability distribution (it gives 'confidence') but it is a distribution function of the parameter, it is a random variable when conditioned on the true parameter value, and unlike a distribution which is coupled with a random variable, it is not unique. I also find a mismatch between mathematical rigor and intuition.
Please explain the definition and assumptions in as much detail as possible (including the arguments of the function and its output) and point me prerequisites or to sources that give more details and illustrative examples. Please explain the motive behind the development of this concept.