Sets of Proofs

Maximal Jaccard Distance

Generalized Clustering Coefficient

• Variant of the original GCC
• Weighted graph
  • Artifact sets as nodes
  • Jaccard similarity as edge weights
  • Triplets, jaw edges and closing edge
• Compute extent of clustering, C_sw =
    ∑ (avg_jaw_weight × closing_weight)
    ------------------------------
          ∑ avg_jaw_weight

• Lower C_sw ≡ greater proof diversity
• Examples
  • P1-P7-P8 = 0.58
  • P2-P3-P4 = 0.26
  • Set of same proofs = 1.00
  • Set of completely different proofs = 0.00