Supplementary Tables 3–8 for the DeFs-CBDE: Clustering-guided binary mutation in multi-objective differential evolution for microarray gene selection

Table 3. The performance comparison of classifiers on the Brain dataset. All Features IG IG - GA IGR IGR - GA CS CS - GA DeFs - CBDE (Best) DeFs - CBDE (Mean ± Std) Accuracy 69.05 73.81 85.71 78.57 97.62 83.33 97.62 100 97.48 ± 3.14 SVM Recall 58 63 75 78 97.5 83 97.5 100 96.25 ± 3.76 Precision 42.28 64.05 68.57 66.52 98.18 88.41 98.18 100 96.58 ± 3.87 F - measure 48.91 63.52 71.64 71.8 97.84 85.62 97.84 100 96.98 ± 3.45 Accuracy 69.05 88.1 92.86 85.71 95.24 83.33 95.24 85.71 82.81 ± 2.14 NB Recall 60.5 80.5 88 78 93 76.5 93 85.71 82.17 ± 2.13 Precision 58.69 90.91 94.55 88.72 95.96 88.77 95.96 78.57 74.59 ± 2.32 F - measure 59.58 85.39 91.16 83.02 94.46 82.18 94.46 80.95 77.12 ± 2.25 Accuracy 78.57 80.95 92.86 83.33 97.62 80.95 97.62 100 95.63 ± 5.40 KNN Recall 75.5 80.5 92.5 83 97.5 80.5 97.5 100 97.90 ± 2.58 Precision 86.36 84.33 94.85 88.33 98.18 87.05 98.18 100 92.07 ± 4.54 F - measure 80.57 82.37 93.66 85.58 97.84 83.65 97.84 100 97.31 ± 2.74 Accuracy 50 61.9 85.71 64.29 88.1 69.05 85.71 85.71 83.46 ± 2.52 DT Recall 48.5 65 87 61.5 89.5 69 83.5 85.71 83.97 ± 2.49 Precision 53.06 61.73 87.22 63.83 88.5 69.89 88.01 78.57 75.99 ± 2.74 F - measure 50.68 63.32 87.11 62.64 89 69.44 85.7 80.95 77.62 ± 3.09 Accuracy 78.57 90.48 100 92.86 100 88.1 100 100 96.02 ± 4.06 RF Recall 76.5 88 100 93 100 88 100 100 96.35 ± 3.57 Precision 80.21 92.73 100 94.36 100 91.33 100 100 94.41 ± 4.52 F - measure 78.31 90.3 100 93.68 100 89.63 100 100 95.35 ± 4.18 Table 4. The performance comparison of classifiers on the Breast dataset. All Features IG IG - GA IGR IGR - GA CS CS - GA DeFs - CBDE (Best) DeFs - CBDE (Mean ± Std) Accuracy 52.58 74.23 84.54 69.07 82.47 73.2 82.47 86.67 84.35 ± 1.92 Recall 50 73.89 84.34 68.67 82.27 73.34 82.48 86.67 84.20 ± 2.03 SVM Precision 26.29 74.41 84.69 69.21 82.6 73.3 82.43 86.67 84.14 ± 2.01 F - measure 34.46 74.15 84.51 68.94 82.43 73.32 82.45 86.67 83.96 ± 2.14 Accuracy 48.45 55.67 57.73 54.64 62.89 72.16 79.38 76 73.10 ± 1.91 NB Recall 46.93 53.47 55.54 52.71 60.87 72.04 79.33 70 67.99 ± 1.72 Precision 45.32 62.94 70.63 56.23 79.31 72.09 79.33 70.09 68.27 ± 1.43 F - measure 46.11 57.82 62.18 54.41 68.88 72.06 79.33 69.97 67.91 ± 1.50 Accuracy 55.67 71.13 89.69 64.95 86.6 72.16 84.54 80 76.00 ± 2.06 KNN Recall 54.43 70.52 89.77 63.58 85.98 71.5 84.44 63.33 60.74 ± 1.65 Precision 55.78 71.93 89.67 69.3 88.89 73.25 84.53 63.56 61.49 ± 1.38 F - measure 55.1 71.22 89.72 66.32 87.41 72.36 84.48 63.37 60.58 ± 1.66 Accuracy 57.73 67.01 86.6 60.82 90.72 69.07 84.54 86.67 84.85 ± 1.47 DT Recall 57.25 66.71 86.3 60.51 90.43 68.88 84.34 90 86.89 ± 1.92 Precision 57.52 66.97 87.09 60.67 91.31 69 84.69 86.67 83.97 ± 1.79 F - measure 57.38 66.84 86.69 60.59 90.87 68.94 84.51 86.79 83.69 ± 1.95 Accuracy 63.92 86.6 89.69 87.63 93.81 81.44 85.57 93.33 91.59 ± 1.42 RF Recall 63.55 86.51 89.66 87.49 93.8 81.29 85.64 94.17 91.58 ± 1.65 Precision 63.85 86.6 89.66 87.71 93.8 81.48 85.54 93.33 90.47 ± 1.80 F - measure 63.7 86.55 89.66 87.6 93.8 81.38 85.59 93.33 89.83 ± 1.91 Table 5. The performance comparison of classifiers on the Lung dataset. All Features IG IG - GA IGR IGR - GA CS CS - GA DeFs - CBDE (Best) DeFs - CBDE (Mean ± Std) Accuracy 78.82 92.12 94.09 83.25 94.58 84.24 95.07 97.56 96.88 ± 0.45 SVM Recall 41.13 71.6 75.68 52.8 86.47 55.15 90.03 97.56 96.62 ± 0.56 Precision 75.27 75.72 76.41 54.47 98.53 54.9 98.66 98.37 97.33 ± 0.57 F - measure 53.19 73.6 76.04 53.62 92.11 55.02 94.15 97.79 96.75 ± 0.61 Accuracy 90.15 95.07 98.52 93.6 97.54 92.12 97.04 96.72 96.06 ± 0.45 NB Recall 79.07 88.5 97.73 93.64 97.44 93.21 97.3 96.72 95.94 ± 0.45 Precision 88.21 94.36 98.54 88.9 93.92 84.92 93.05 95.23 94.39 ± 0.54 F - measure 83.39 91.34 98.13 91.21 95.65 88.87 95.13 95.08 94.27 ± 0.53 Accuracy 92.61 92.61 97.04 89.66 96.06 92.12 95.57 96.72 96.27 ± 0.33 KNN Recall 80.73 87.91 94.87 73.98 92.74 80.36 91.79 96.72 96.16 ± 0.35 Precision 95.22 89.1 95.4 93.1 97.78 95.09 97.65 97.20 96.57 ± 0.40 F - measure 87.38 88.5 95.13 82.45 95.19 87.11 94.63 96.81 96.24 ± 0.35 Accuracy 84.73 86.7 96.55 84.73 96.06 85.22 96.55 97.56 97.27 ± 0.21 DT Recall 69.07 73.69 94.68 69.07 92.2 72.4 93.15 97.56 97.20 ± 0.25 Precision 84.15 80.63 94.35 84.15 95.21 85.92 96.21 97.63 97.33 ± 0.22 F - measure 75.87 77 94.51 75.87 93.68 78.58 94.66 97.17 96.90 ± 0.23 Accuracy 83.74 93.6 96.06 91.13 95.57 92.61 96.06 91.80 91.51 ± 0.22 RF Recall 59.1 85.6 90.58 78.46 92.01 83.47 92.97 91.80 91.42 ± 0.24 Precision 93.54 97.15 97.86 94.45 97.73 96.82 97.86 92.73 92.34 ± 0.20 F - measure 72.43 91.01 94.08 85.72 94.78 89.65 95.35 90.65 90.28 ± 0.21 Table 6. The performance comparison of classifiers on the CNS dataset. All Features IG IG - GA IGR IGR - GA CS CS - GA DeFs - CBDE (Best) DeFs - CBDE (Mean ± Std) Accuracy 65 65 86.67 65 65 65 83.33 91.67 91.37 ± 0.16 SVM Recall 50 50 82.05 50 50 50 77.29 91.67 91.36 ± 0.17 Precision 32.5 32.5 88.89 32.5 32.5 32.5 86.58 92.59 92.31 ± 0.13 F - measure 39.39 39.39 85.33 39.39 39.39 39.39 81.67 91.32 91.06 ± 0.12 Accuracy 61.67 75 90 78.33 88.33 70 83.33 88.89 88.59 ± 0.14 NB Recall 59.52 74.18 87.91 75.64 87.73 69.23 81.68 88.89 88.52 ± 0.14 Precision 59.03 72.92 89.86 76.25 86.96 68 81.68 92.59 92.19 ± 0.13 F - measure 59.27 73.54 88.87 75.94 87.34 68.61 81.68 89.57 89.18 ± 0.14 Accuracy 61.67 75 93.33 68.33 83.33 75 88.33 83.33 83.00 ± 0.17 KNN Recall 54.03 71.98 91.58 61.36 78.39 69.78 84.43 83.33 82.92 ± 0.15 Precision 55.12 72.5 93.71 64.44 84.44 73.01 90.06 86.46 85.89 ± 0.19 F - measure 54.57 72.24 92.63 62.86 81.3 71.36 87.15 82.43 82.00 ± 0.17 Accuracy 58.33 70 93.33 68.33 93.33 61.67 88.33 72.22 71.72 ± 0.21 DT Recall 54.76 67.03 91.58 63.55 91.58 54.03 84.43 72.22 71.62 ± 0.23 Precision 54.67 67.03 93.71 64.68 93.71 55.12 90.06 79.17 78.49 ± 0.21 F - measure 54.71 67.03 92.63 64.11 92.63 54.57 87.15 72.49 71.86 ± 0.21 Accuracy 53.33 80 91.67 80 90 83.33 88.33 77.78 77.24 ± 0.22 RF Recall 45.42 75.82 89.19 72.53 85.71 78.39 85.53 77.78 77.05 ± 0.26 Precision 44.44 78.93 92.46 72.53 93.33 84.44 88.49 82.72 81.94 ± 0.29 F - measure 44.92 77.34 90.8 72.53 89.36 81.3 86.98 68.06 66.94 ± 0.61 Table 7. Evaluation of DeFs-CBDE through comparison with other existing feature selection methods in terms of best classification accuracy. Dataset Classifier Original dataset (without FS) S Prajapati et al. H Hamla and K Ghanem Ahadzadeh et al DeFs-CBDE Brain SVM 69 .05 / / / 100 NB 69 .05 / / / 85.71 KNN 78.57 / / / 100 DT 50.00 / / / 85.71 RF 78.57 / / / 100 Breast SVM 52.58 / 100 / 86.67 NB 48.45 / 69.36 / 76.00 KNN 55.67 / 85.95 94.94 80.00 DT 57.73 70 72.61 / 86.67 RF 63.92 75 84.09 / 93.33 Lung SVM 78.82 / 100 / 97.56 NB 90.15 / 92.12 / 96.72 KNN 92.61 / 92.56 99.02 96.72 DT 84.73 88.51 91.72 / 97.56 RF 83.74 87.80 91.72 / 91.80 CNS SVM 65.00 / 100 / 91.67 NB 61.67 / 85.50 / 88.89 KNN 61.67 / 90.33 96.66 83.33 DT 58.33 83.33 77.83 / 72.22 RF 53.33 66.66 86.33 / 77.78 Table 8. Enrichment results of FS-selected genes (Lung dataset, g:Profiler). ID Source Term ID Term Name P_adj (query_1) 1 GO:MF GO:0005515 protein binding 4.740×10⁻⁷⁰ 2 GO:MF GO:0016301 kinase activity 8.162×10⁻⁸ 3 GO:MF GO:0001216 DNA-binding transcription activator activity 2.369×10⁻ 6 4 GO:MF GO:0003690 double-stranded DNA binding 1.165×10⁻⁵ 5 GO:MF GO:0035005 1-phosphatidylinositol-4-phosphate 3-kinase activity 5.772×10⁻⁴ 6 GO:MF GO:0042578 phosphoric ester hydrolase activity 2.255×10⁻² 7 GO:MF GO:0017111 ribonucleoside triphosphate phosphatase activity 2.316×10⁻² 8 GO:MF GO:0016307 1-phosphatidylinositol-3-kinase activity 2.535×10⁻² 9 GO:MF GO:0004016 adenylate cyclase activity 4.119×10⁻² 10 GO:BP GO:0032501 multicellular organismal process 1.097×10⁻ 89 11 GO:BP GO:0016054 organic acid catabolic process 2.380×10⁻⁹ 12 GO:BP GO:0019318 phosphatidylinositol-3-phosphate biosynthetic process 2.387×10⁻⁸ 13 GO:BP GO:0007601 visual perception 2.105×10⁻⁵ 14 GO:BP GO:0022414 reproductive process 5.497×10⁻⁴ 15 GO:BP GO:0072521 purine-containing compound metabolic process 7.213×10⁻⁴ 16 GO:BP GO:0019313 carbohydrate derivative metabolic process 1.307×10⁻³ 17 GO:BP GO:0042354 L-fucose metabolic process 6.232×10⁻³ 18 GO:BP GO:0007049 cell division 7.033×10⁻³ 19 GO:BP GO:0009308 amine metabolic process 1.328×10⁻² 20 GO:BP GO:0048545 rhythmic process 1.773×10⁻² 21 GO:BP GO:0007612 cell recognition 2.023×10⁻² 22 GO:BP GO:0000793 supramolecular fiber organization 2.120×10⁻² 23 GO:BP GO:0046068 cGMP metabolic process 2.396×10⁻² 24 GO:BP GO:0003085 regulation of systemic arterial blood pressure by hormone 2.491×10⁻² 25 GO:BP GO:0009749 response to glucose 4.196×10⁻² 26 GO:CC GO:0005813 cell periphery 6.517×10⁻ 57 27 GO:CC GO:0022626 cytosolic ribosome 7.834×10⁻⁸ 28 GO:CC GO:0000785 chromatin 4.511×10⁻⁵ 29 GO:CC GO:0005856 cytoskeleton 5.061×10⁻⁵ 30 GO:CC GO:0005813 centrosome 1.965×10⁻²