Algorithm 2.
Empirical Data Creation (D, n, r)
| Input: dataset D, number of bootstraps n, and a set of constraints r | |
| Output: empirical dataset Dr with n rows and m = |r| columns | |
| 1: | Let Dr[n, m] be a new 2−d array with n rows and m columns |
| 2: | for b = 1 to n do |
| 3: | sampleb ← Bootstrap(D) |
| 4: | for ri ∈ {r1, r2, …, rm} do |
| 5: | p ← BSC(ri, sampleb) |
| 6: | if p ≥ 0.5 then |
| 7: | Dr[b, i] ← 1 |
| 8: | else |
| 9: | Dr[b, i] ← 0 |
| 10: | return Dr[n, m] |