Table 5 Performance of applied combinations of dimensionality reduction and classification techniques on test set for CBR and CBREXPFIN data.
From: Decision making on vestibular schwannoma treatment: predictions based on machine-learning analysis
Feature selection method | Classification accuracy (ACC) for test set (%) | ||||||
|---|---|---|---|---|---|---|---|
Decision tree | Rand. Forest | Gradient boosting | Logistic regression | Supp. vect. machine | Neural network | Avg | |
Decision tree (CBR) | 76 | 79 | 84 | 82 | 79 | 76 | 79 |
Random forest (CBR) | 76 | 79 | 87 | 76 | 76 | 68 | 77 |
Gradient boosting (CBR) | 84 | 82 | 86 | 84 | 76 | 89 | 84 |
Logistic regression (CBR) | 67 | 78 | 68 | 86 | 83 | 85 | 78 |
LASSO (CBR) | 74 | 79 | 79 | 76 | 74 | 82 | 77 |
Expert (CBREXPFIN) | 76 | 76 | 87 | 79 | 76 | 74 | 78 |
Avg | 76 | 79 | 82 | 81 | 77 | 79 | 79 |
