Figure 3From: Exploiting likely-positive and unlabeled data to improve the identification of protein-protein interaction articlesImpact of adding likely data on different term weighting schemes (F-measure).Back to article page