The impact of chemotherapy and survival prediction by machine learning in early Elderly Triple Negative Breast Cancer (eTNBC): a population based study from the SEER database

Table 6 Model performance for 5-year BCSS and 5-year OS

Algorithms	Accuracy	Precision	Sensitivity	F1 score	AUC
5-year BCSS
K-nearest neighbor	0.879	0.882	0.98	0.928	0.70
Catboost	0.905	0.892	0.974	0.932	0.69
Decision tree	0.908	0.901	0.949	0.924	0.61
Random forest	0.869	0.889	0.971	0.929	0.70
Gradient booster	0.882	0.887	0.991	0.936	0.75
LightGBM	0.882	0.887	0.991	0.936	0.75
Neural network model	0.886	0.877	1.0	0.934	0.75
Support vector machine	0.882	0.887	0.991	0.936	0.51
XGBoost	0.879	0.892	0.98	0.934	0.70
5-year OS
K-nearest neighbor	0.844	0.857	0.952	0.902	0.73
Catboost	0.877	0.86	0.977	0.915	0.76
Decision tree	0.882	0.869	0.940	0.903	0.69
Random forest	0.837	0.864	0.954	0.907	0.72
Gradient booster	0.849	0.855	0.985	0.916	0.80
LightGBM	0.851	0.859	0.983	0.916	0.81
Neural network model	0.86	0.877	0.949	0.911	0.79
Support vector machine	0.854	0.854	0.994	0.919	0.70
XGBoost	0.865	0.868	0.988	0.924	0.79

ISSN: 1471-2318