Table 5.

Performance of classification model for prediction of response to therapy in for urothelial bladder cancer patients using different machine-learning classifiers

Classifier	Hyperparameter	Accuracy (%)	Sensitivity	Specificity
Logistic regression	Penalty: l2 C: 0.0464	85.5 ± 5.2	82.5	93.03
Decision tree	Best criterion: entropy Best max_depth: 6	88.5 ± 3.4	91.2	83.7
K-nearest neighbor	Algorithm: auto leaf_size: 30, metric: minkowski n_jobs: None n_neighbors: 10	90.0 ± 4.5	87.7	90.7
Support vector machine	Kernel: rbf C: 100 gamma: 0.01	86.0 ± 4.9	92.98	81.4
Random Forest	max_depth: 5 min_samples_leaf: 4 n_estimators: 100	89.5 ± 3.7	87.7	83.7
Voting	KNN + DT + RF	90.0 ± 3.4	92.98	81.4