Algorithm 6.

DFG Algorithm

Input:
	Graph database D; Threshold θ; Privacy budgets ϵ1, ϵ2 and ϵ3 (ϵ1 + ϵ2 + ϵ3 = ϵ);
Output:
	A set F of frequent subgraphs together with their noisy supports;
1:	ζ ← get the maximum support of subgraphs with different sizes; \\ this step can be done offline
2:	Mg ← get the maximum frequent subgraph size based on ζ using ϵ1;
3:	F ← frequent_subgraph_identification(D, θ, ϵ2, Mg);
4:	lattice-based_noisy_support_computation(F, ϵ3);
5:	return F;