Research Article
Joint Nonnegative Matrix Factorization Based on Sparse and Graph Laplacian Regularization for Clustering and Co-Differential Expression Genes Analysis
Table 3
Performance of different analysis methods.
| | | Datasets | iNMF | iGMFNA | jNMF | iONMF | SG-jNMF |
| | Accuracy | PAAD | 53.56 (0.00) | 63.44 (3.05) | 53.11 (0.01) | 56.23 (0.72) | 95.00 (0.00) | | CHOL | 90.22 (1.71) | 97.33 (0.04) | 93.78 (0.49) | 90.04 (0.96) | 99.11 (0.00) | | ESCA | 54.17 (0.00) | 54.58 (0.00) | 53.96 (0.00) | 58.70 (1.46) | 66.87 (0.01) | | COAD | 61.87 (0.01) | 63.35 (0.02) | 59.15 (0.09) | 62.13 (0.22) | 68.84 (0.01) |
| | Recall | PAAD | 51.47 (4.68) | 58.07 (4.84) | 53.34 (4.50) | 48.44 (1.20) | 67.33 (2.26) | | CHOL | 50.78 (2.01) | 56.17 (1.84) | 55.39 (2.04) | 46.41 (1.49) | 88.06 (0.47) | | ESCA | 50.98 (0.29) | 51.31 (0.29) | 48.86 (0.27) | 51.13 (0.30) | 51.04 (0.12) | | COAD | 49.34 (1.25) | 49.15 (1.30) | 49.65 (1.28) | 46.73 (1.08) | 56.18 (1.51) |
| | Precision | PAAD | 97.79 (0.05) | 98.60 (0.04) | 97.52 (0.05) | 97.71 (0.05) | 99.09 (0.01) | | CHOL | 58.55 (0.20) | 62.21 (0.20) | 63.36 (0.21) | 60.54 (1.72) | 91.00 (0.10) | | ESCA | 94.40 (0.05) | 95.20 (0.07) | 95.90 (0.08) | 95.67 (0.06) | 98.39 (0.02) | | COAD | 49.34 (1.25) | 49.15 (1.30) | 49.65 (1.28) | 46.73 (1.08) | 56.18 (0.51) |
| | F1-score | PAAD | 63.54 (3.70) | 66.70 (1.29) | 67.84 (1.71) | 64.42 (1.86) | 77.15 (2.12) | | CHOL | 53.47 (2.01) | 58.19 (1.91) | 58.03 (2.04) | 51.61 (1.63) | 89.18 (1.04) | | ESCA | 66.00 (0.18) | 66.47 (0.19) | 64.52 (0.18) | 66.41 (0.16) | 66.96 (0.16) | | COAD | 63.71 (1.34) | 63.56 (1.27) | 64.25 (1.30) | 61.23 (1.14) | 71.11 (0.55) |
|
|
