Peritoneal cytology predicting distant metastasis in uterine carcinosarcoma: machine learning model development and validation

Table 5 Model Performance Index Evaluation

Evaluation of the performance of the training set
Model	Threshold	Accuracy	Sensitivity	Specificity	Precision	F1
Logistic	0.113	0.768	0.914	0.733	0.456	0.609
SVM	0.146	0.779	0.882	0.754	0.468	0.612
GBM	0.189	0.814	0.890	0.796	0.517	0.654
NeuralNetwork	0.239	0.817	0.831	0.813	0.522	0.641
RandomForest	0.500	0.972	0.867	0.998	0.990	0.925
KNN	0.202	0.831	0.975	0.796	0.540	0.695
Adaboost	0.500	0.852	0.622	0.908	0.625	0.624
LightGBM	0.166	0.791	0.905	0.763	0.484	0.631
Test set performance evaluation
Model	Threshold	Accuracy	Sensitivity	Specificity	Precision	F1
Logistic	0.113	0.758	0.872	0.730	0.442	0.587
SVM	0.146	0.779	0.862	0.758	0.467	0.606
GBM	0.189	0.805	0.842	0.796	0.503	0.630
NeuralNetwork	0.239	0.806	0.788	0.810	0.505	0.615
RandomForest	0.500	0.816	0.458	0.903	0.538	0.495
KNN	0.202	0.793	0.833	0.784	0.486	0.613
Adaboost	0.500	0.860	0.606	0.923	0.658	0.631
LightGBM	0.166	0.769	0.808	0.759	0.452	0.580

ISSN: 1477-7819