Skip to main content
Fig. 1 | BMC Bioinformatics

Fig. 1

From: Leveraging permutation testing to assess confidence in positive-unlabeled learning applied to high-dimensional biological datasets

Fig. 1

PU learning use cases, synthetic data generation, data unlabeling. A Examples of settings in which PU learning may improve class prediction. B Synthetic datasets in which there is no (distance d = 0), low (d = 1), or high (d = 2) degree of difference between positive (P) and negative (N) classes were generated. C The unlabeled (U) sample set was comprised of varying proportions (10%, 30%, 50%) of N among U samples. D Exemplary visualization of P and N sample data profiles. E Exemplary partial relabeling of P/N class data to generate P and U classes

Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us