Table 3 Performance comparison of three machine learning models.

From: Utilizing SMOTE-TomekLink and machine learning to construct a predictive model for elderly medical and daily care services demand

Category

Model

ACC

R

P

F1-score

AUC

Demand for medical services

RF

0.843

0.862

0.973

0.914

0.650

GBDT

0.787

0.928

0.626

0.747

0.853

LightGBM

0.846

0.870

0.814

0.841

0.910

Demand for daily care services

RF

0.763

0.803

0.884

0.841

0.819

GBDT

0.794

0.867

0.648

0.741

0.848

LightGBM

0.830

0.867

0.777

0.819

0.906

Back to article page