Role of k in Delta^k: Interaction Order Sensitivity

Computational investigation of the conjecture that k tunes order sensitivity

Definition

Delta^k(X) = (N - k) * T(X) - Sum_i T(X_{-i})

Where T(X) is the total correlation and X_{-i} is the system with variable i removed. Special cases: k=0 gives S-information, k=1 gives dual total correlation, k=2 gives -O-information.

Configuration

5

Sensitivity Profile |Delta^k| vs Order m

Special Cases Comparison

Blind Spot Verification (Delta^k at m=k)

Delta^k Across All Orders