Table 2.
Method | MCC | Accuracy | Sensitivity | Specificity | AUC | P-value |
---|---|---|---|---|---|---|
AIPpred | 0.479 | 0.744 | 0.741 | 0.746 | 0.814 | |
ERT | 0.463 | 0.736 | 0.731 | 0.740 | 0.804 | 0.621 |
AntiInflam (MA) | 0.210 | 0.601 | 0.786 | 0.417 | 0.706 | <0.000001 |
SVM | 0.298 | 0.651 | 0.621 | 0.680 | 0.704 | <0.000001 |
k-NN | 0.296 | 0.640 | 0.479 | 0.801 | 0.699 | <0.000001 |
AntiInflam (LA) | 0.197 | 0.575 | 0.258 | 0.892 | 0.647 | <0.000001 |
The first column represents the method employed in this study. The second, the third, the fourth, and the fifth respectively represent the MCC, accuracy, sensitivity, and specificity. The sixth column and the seventh represent the AUC and pairwise comparison of AUC between AIPpred and the other methods computed using a two-tailed t-test. A P ≤ 0.05 indicates a statistically meaningful difference between AIPpred and the selected method (shown in bold). In the first column, LA and MA respectively correspond to less accurate and more accurate prediction method. We note that AntiInflam LA and MA classification accuracy was computed using default threshold value of 0.5 and -0.3 (reported in Gupta et al., 2017b), respectively.