Figure 4

The scatter plots between F1 score and the likelihood functions. The upper row is the relationships between the sum of similarity (LS) and F1 score and, in the bottom row, the product of similarity (LP) and F1 score are depicted using compound standards homogeneous data (first column), spiked-in homogeneous data (second column), and compound standards heterogeneous data (last column). In each plot, the black circle and solid line are for SWRM method, the red triangle and dashed line for SWRE method, and the green plus ('+') and dotted line for SWRME method. The lines represent the linear regression fitting.