Table 3 Accuracy results of thirteen methods on four testing sets.
From: Heterogeneity-optimized method for predicting immune checkpoint blockade response
Pan-cancer | Melanoma | NSCLC | Other cancers | |
|---|---|---|---|---|
Our model | 78.24% | 76.52% | 80.48% | 81.15% |
RF16 | 73.22% | 56.76% | 74.70% | 73.88% |
RF11 | 72.88% | 64.86% | 72.78% | 72.15% |
TMB | 63.17% | 45.78% | 61.75% | 56.18% |
SVM | 71.88% | 67.57% | 73.15% | 63.15% |
MLR | 73.89% | 71.22% | 70.85% | 66.85% |
NB | 73.00% | 65.22% | 63.98% | 70.15% |
GBDT | 73.00% | 55.85% | 74.77% | 68.67% |
AdaBoost | 74.21% | 72.84% | 76.64% | 72.53% |
LightGBM | 77.00% | 75.43% | 78.98% | 73.94% |
ExtraTrees | 71.86% | 65.57% | 78.51% | 74.57% |
BPNN | 69.49% | 56.47% | 76.64% | 70.67% |
DNN | 63.72% | 45.68% | 75.64% | 70.00% |
