Table 5 Accuracy of three-class classification in 1D and 2D modes.
From: Stochastic differential equation modeling approach for grading astrocytomas on brain MRI images
Processing Mode | Classifier | Accuracy (mean ± std) |
|---|---|---|
Two-dimensional (2D) | RBF-SVM | 98.49% ± 0.9% |
KNN (k = 3, Cityblock distance) | 98.45% ± 0.6% | |
KNN (k = 3, Euclidean distance) | 98.22% ± 0.3% | |
KNN (k = 3, Mahalanobis distance) | 97.13% ± 1.2% | |
RF | 95.42% ± 1.7% | |
One-dimensional (1D) | RBF-SVM | 93.52% ± 1.4% |
KNN (k = 3, Euclidean distance) | 92.28% ± 2.0% | |
KNN (k = 3, Mahalanobis distance) | 92.13% ± 1.2% | |
KNN (k = 3, Cityblock distance) | 91.81% ± 1.3% | |
RF | 87.20% ± 2.1% |
