An Iterative Information-Theoretic Approach to the Detection of Structures in Complex Systems

Table 3

Test case: RSs found in the main iterations of the sieving algorithm and corresponding values with . At the end of each iteration, the RS with the highest is grouped into a single variable for the next iteration. In iteration 1, we show also the second RS ranked. Note that, in iteration 3, the of group ( ) (lower than the chosen threshold of 3.0) makes it not relevant.

Synthetic data 1: independent cliques ()
Iteration	Relevant set(s)
1	⟶   ()	1517.01
		1454.52
2	⟶   ()	684.923
3		1.644