Table 10 Performance of applied combinations of dimensionality reduction and classification techniques on test set for PDA and PDAEXPFIN 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 (PDA) | 80 | 80 | 80 | 90 | 80 | 70 | 80 |
Random forest (PDA) | 80 | 70 | 80 | 70 | 90 | 70 | 77 |
Gradient boosting (PDA | 80 | 80 | 80 | 90 | 80 | 80 | 82 |
Logistic regression (PDA) | 80 | 60 | 80 | 80 | 90 | 90 | 80 |
LASSO (PDA) | 90 | 90 | 90 | 90 | 80 | 90 | 88 |
Expert (PDAEXPFIN) | 79 | 79 | 79 | 63 | 63 | 68 | 72 |
Avg | 82 | 77 | 82 | 81 | 81 | 78 | 80 |
