Table 10 Best performing classifications schemes based on each source of data
From: A multiscale and multiparametric approach for modeling the progression of oral cancer
Source of data | Feature selection | Classifier | Acc (%) | Se (%) | Sp (%) | Kappa | AUC |
|---|---|---|---|---|---|---|---|
Clinical | Wrapper | DT | 83.3 (±10.06) | 73.7 (±14.35) | 93 (±8.05) | 0.67 (±0.20) | 0.842 (±0.09) |
Imaging | Wrapper | NB | 90.9 (±12.25) | 88.6 (±14.76) | 93.2 (±14.15) | 0.82 (±0.24) | 0.89 (±0.12) |
Tissue genomic | Union of genes from literature and current work | ANN | 91.23 (±7.23) | 94.7 (±11.25) | 87.7 (±13.72) | 0.82 (±0.14) | 0.957 (±0.04) |
Blood genomic | No feature selection | ANN | 95.8 (±15.81) | 100 (±0) | 91.7 (±31.62) | 0.92 (±0.32) | 1 (±0) |