A Multicriteria Approach to Find Predictive and Sparse Models with Stable Feature Selection for High-Dimensional Data

Table 6

Stability values for the configurations in Table 5. The ID column references Table 5. “—” means that the configuration is not Pareto optimal for the corresponding stability measure.

ID	1 − SJ	1 − SJ	1 − SC	1 − SC	1 − SD-1	1 − SD-1	1 − SD-10	1 − SD-10
	(train)	(test)	(train)	(test)	(train)	(test)	(train)	(test)
1	0.039	0.046	0.021	0.025	—	—	—	—
2	0.040	0.054	0.021	0.029	—	—	—	—
3	0.041	0.060	0.021	0.032	0.062	0.091	—	—
4	0.047	0.060	0.025	0.032	0.066	0.097	—	—
5	0.063	0.054	0.034	0.028	0.081	0.073	—	—
6	0.072	0.057	0.038	0.030	0.100	0.090	0.243	0.234
7	0.178	0.340	0.100	0.208	—	—	—	—
8	0.206	0.283	0.117	0.169	0.247	0.447	0.255	0.453
9	0.482	0.503	0.325	0.341	0.597	0.561	0.605	0.570
10	0.345	0.284	0.240	0.195	0.455	0.441	1.000	1.000
11	0.131	0.158	0.075	0.092	0.253	0.277	0.879	0.903