Table 3.
Comparative analysis of the grid search optimization of selected ML algorithms with SS, SP and RFE pre-processing operators for P1(A) and P2(B)
| Model | Balanced accuracy V/T | Precision V/T | Recall V/T | ROC AUC V/T | PRC V/T | Pipeline complexity |
|---|---|---|---|---|---|---|
| A. P1 | ||||||
| LR pipelines | ||||||
| LR | 0.68/0.72 | 0.85/0.87 | 0.85/0.90 | 0.68/0.87 | 0.84/0.95 | 1 |
| LR +SS | 0.68/0.77 | 0.86/0.92 | 0.76/0.80 | 0.68/0.84 | 0.84/0.94 | 2 |
| LR + SS + SP | 0.68/0.78 | 0.86/0.92 | 0.76/0.81 | 0.68/0.84 | 0.84/0.94 | 3 |
| LR + SS + RFE | 0.73/0.77 | 0.88/0.91 | 0.80/0.81 | 0.73/0.84 | 0.86/0.94 | 3 |
| BNB pipelines | ||||||
| BNB | 0.72/0.77 | 0.88/0.91 | 0.76/0.78 | 0.72/0.85 | 0.85/0.95 | 1 |
| BNB + SS | 0.66/0.72 | 0.85/0.89 | 0.73/0.73 | 0.66/0.79 | 0.83/0.92 | 2 |
| BNB + SS + SP | 0.71/0.76 | 0.88/0.92 | 0.76/0.75 | 0.71/0.84 | 0.85/0.95 | 3 |
| BNB + SS + RFE | 0.70/0.73 | 0.88/0.90 | 0.75/0.76 | 0.70/0.82 | 0.85/0.94 | 3 |
| B. P2 | ||||||
| LR pipelines | ||||||
| LR | 0.73/0.74 | 0.76/0.78 | 0.84/0.84 | 0.73/0.85 | 0.73/0.89 | 1 |
| LR +SS | 0.69/0.76 | 0.73/0.81 | 0.79/0.80 | 0.69/0.84 | 0.70/0.89 | 2 |
| LR + SS + SP | 0.72/0.76 | 0.76/0.81 | 0.80/0.79 | 0.72/0.83 | 0.73/0.89 | 3 |
| LR + SS + RFE | 0.74/0.74 | 0.78/0.80 | 0.81/0.80 | 0.74/0.84 | 0.74/0.89 | 3 |
| BNB pipelines | ||||||
| BNB | 0.70/0.74 | 0.75/0.80 | 0.77/0.79 | 0.70/0.80 | 0.71/0.85 | 1 |
| BNB + SS | 0.63/0.66 | 0.69/0.75 | 0.69/0.67 | 0.63/0.74 | 0.66/0.81 | 2 |
| BNB + SS + SP | 0.74/0.75 | 0.80/0.83 | 0.74/0.75 | 0.74/0.82 | 0.74/0.87 | 3 |
| BNB + SS + RFE | 0.65/0.67 | 0.71/0.76 | 0.74/0.73 | 0.65/0.77 | 0.68/0.83 | 3 |
Note : SS, standard scaler; SP, select percentile; RFE, recursive feature eliminator; LR, logistic regression classifier; BNB, Bernoulli Naïve Bayes classifier.