A 2-means clustering step cleans labels to enable logistic regression for positive-unlabeled classification without requiring the SCAR assumption.
A vailable from http:arxiv.org/abs/1811.04820v3 (2020) A proposal for PU
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ME 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A proposal for PU classification under Non-SCAR using clustering and logistic model
A 2-means clustering step cleans labels to enable logistic regression for positive-unlabeled classification without requiring the SCAR assumption.