Table 4 The performance of the decision tree, random forests, artificial neural networks, and support vector machine models.
From: Performance and efficiency of machine learning algorithms for analyzing rectangular biomedical data
Statistical Measure | Alive (%) | Non-breast cancer (%) | Breast cancer (%) | CVS (%) | Other cause (%) |
|---|---|---|---|---|---|
Decision tree | |||||
Accuracy | 69.21 | ||||
Precision | 73.56 | NA | 55.65 | 22.58 | 21.09 |
Recall | 94.17 | 0.00 | 43.36 | 2.25 | 8.45 |
Specificity (TNR) | 26.50 | 100.00 | 91.65 | 99.23 | 96.38 |
F1 | 82.60 | NA | 48.74 | 4.09 | 12.06 |
Random forest | |||||
Accuracy | 70.23 | ||||
Precision | 71.56 | 0.00 | 62.55 | 33.33 | 33.33 |
Recall | 96.69 | 0.00 | 43.22 | 2.25 | 1.09 |
Specificity (TNR) | 22.65 | 99.94 | 93.71 | 99.56 | 99.75 |
F1 | 82.25 | NA | 51.12 | 4.22 | 2.11 |
Artificial neural networks | |||||
Accuracy | 70.16 | ||||
Precision | 72.88 | NA | 59.52 | 36.00 | 28.57 |
Recall | 95.47 | 0.00 | 44.04 | 5.79 | 5.45 |
Specificity (TNR) | 25.84 | 100.00 | 92.78 | 98.99 | 98.43 |
F1 | 82.66 | NA | 50.62 | 9.97 | 9.15 |
Support vector machine | |||||
Accuracy | 69.06 | ||||
Precision | 70.04 | NA | 60.75 | NA | NA |
Recall | 96.22 | 0.00 | 39.43 | 0.00 | 0.00 |
Specificity (TNR) | 19.43 | 100.00 | 93.75 | 100.00 | 100.00 |
F1 | 81.07 | NA | 47.82 | NA | NA |
Multinomial logistic regression | |||||
Accuracy | 68.12 | ||||
Precision | 69.71 | 61.10 | 13.73 | 50.00 | 22.73 |
Recall | 96.38 | 42.66 | 1.13 | 0.54 | 1.24 |
Specificity (TNR) | 29.33 | 73.98 | 85.71 | 90.41 | 99.59 |
F1 | 80.90 | 50.25 | 2.10 | 1.07 | 2.35 |
