Cost does not mean class.
In this work, for measuring vessel classification performance, we report mean per class accuracies.
The model performance was compared by evaluating class accuracy, class precision, and kappa scores.
Class accuracy refers to the proportion of a given observed class that was correctly classified.
The gains in class accuracy and class precision for the high vulnerability class with the use of a lower concentration threshold were much greater than the decreases in class accuracy and class precision in the low vulnerability class.
Inhibitors displaying the highest and lowest mean accuracy, respectively by class, in Table 3 of our study are ritonavir (82.7%) and atazanavir (69.3%) for the PIs; emtricitabine (89.4%) and tenofovir (69.6%) for the NRTIs; and nevirapine (81.5%) and delavirdine (74.8%) for the NNRTIs.
G-mean computes the geometric mean of all classes' accuracies and is described as follows: (14) G -mean = (∏ i = 1 C Ac c i ) 1 / C, where Acc i denotes the accuracy of the ith class.
Percent change in 2-back accuracy was calculated as (mean accuracy on 2-back/4 minus mean accuracy on 2-back/1) divided by mean accuracy on 2-back/1.
One interesting phenomenon is that higher genotype imputation accuracy does not necessarily mean higher GEBV accuracy.
Secondly, class accuracies of especially the coniferous and mixed forest classes were enhanced after correction.
