Figure 8
From: Integrating phenotype and gene expression data for predicting gene function
F-measure. The F-measure tends to favor cutoffs which are neither very high nor very low. F-measure is calculated as (2*precision*recall)/(precision + recall) [16]. Thus, the best F-measures strike a balance between precision and recall. The integrated data set using the max to combine the similarities results in the highest F-measure with a cutoff of 0.6. Figure 8a shows the results for the generalized predictions. Figure 8b shows the results for the exact predictions.