Skip to main content
Fig. 3 | BioData Mining

Fig. 3

From: A maximum flow-based network approach for identification of stable noncoding biomarkers associated with the multigenic neurological condition, autism

Fig. 3

Feature stability metrics on simulated data. The graphs above show that the implementation of the maximum flow algorithm improves feature stability across various levels of feature correlation and dataset sizes. The maximum flow procedure also maintains prediction performance (AUC-ROC). Since the Jaccard index is trivially 1.0 after maximum flow, these values have been omitted from the figure

Back to article page

BioData Mining

ISSN: 1756-0381