Table 2.
Given a sample of n individuals with their corresponding score sets X ≡ {x1, x2,…, xn}, each having m different scores, x ≡ {x1,x2,…, xm}, their diagnosis labels y ∈ {0,1}, and noise variances of scores, |
1. For each score sj
|
2. Calculate probabilities p(sj) by normalizing weights wj |
3. Fuse individual severity beliefs to get (Equation 1) |
4. Repeat 1–3 to get a distribution over |
5. Calculate expected value by averaging |