Table 9.
Performance metrics obtained with our proposed framework and other classifiers obtained from the literature for WBCD.
| References | DDFCDSS—This Work | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Main Aspects | [29] | [11] | [10] | [9] | [94] | [6] | [26] | [95] | Five Variables (RS) | Five Variables (CV) | |
| Num of Variables | 3 | 9 | - | 6 | 9 | 9 | 7 | 3 | 5 | 5 | |
| Num of Rules or Hidden neurons/technique | 39/FIS | DBN/4:2 | DNN | DNN | SMO | FCLF and CNN | FRNN | WT and IT2FLS | 192/DDFCDSS | 233/DDFCDSS | |
| Performance | Accuracy (%): | 98.57% | 99.68% | 98.62% | 96.2%:96.6% | 72.70% | 98.71% | 99.72% | 97.88% | 99.28% | 99.40% |
| Sensitivity: | 0.9793 | 1.000 | - | - | - | 0.976 | 1.000 | 0.9850 | 0.9836 | 0.988 | |
| Specificity: | 0.9891 | 0.9947 | - | - | - | 0.9943 | 0.9947 | 0.9650 | 0.9978 | 0.998 | |
| F-Measure: | 0.9793 | - | - | - | 0.71 | - | 0.9970 | - | 0.9897 | 0.992 | |
| Area under curve: | 0.9901 | - | - | - | 0.63 | 0.9816 | 1.000 | 0.9750 | 0.9936 | 0.995 | |
| Kappa statistics: | 0.9683 | - | - | - | - | - | 0.9943 | - | 0.9842 | 0.987 | |
FIS: Fuzzy Inference System. DBN: Deep Belief Network. DNN: Deep Neural Network. SMO: Sequential Minimal Optimization. CNN: Convolutional Neural Network. FCLF: Fully Connected Layer First. FRNN: Fuzzy-Rough Nearest Neighbor. WT: Wavelet transformation. IT2FLS: interval type-2 fuzzy logic system. DDFCDSS: Data-driven Fuzzy Decision Support System. RS: Random Sampling. CV: Cross Validation. - which are not mentioned in the literature. Bold values indicate the best performance with fewer input variables.