Modeling Categorical Variables by Mutual Information Decomposition.

It is possible that two candidates of a new predictor X t + 1 (for some t ≥ 1 ) exist with two close (significant) MIR estimates, or two CMI estimates with closely significant p values.

also in 33 other papers

Next, the second summand of (21) indicated that the factor M (smoking) was marginally significant conditional on the other four factors.

also in 26,082 other papers

Note that the graphs were constructed from models comprising mainly significant two-way effects (edges) between the variables (nodes), with possibly significant cliques (triangles).

also in 649 other papers