Detecting High-Risk Factors and Early Diagnosis of Diabetes Using Machine Learning Methods

Table 3

Comparison of the proposed method with existing studies that used other datasets.

Study	Dataset	Method	Accuracy (%)	Precision	Sensitivity	Specificity	F-measure
[8]	Private	RF	80.84	—	0.85	0.767	—
	PIDD	RF	77.21		0.746	0.799
[13]	PIDD	SVM, AB	94.44	0.971	0.910	—	—
[16]	PIDD	LR,SVM	78.85, 77.71	0.788, 0.774	0.789, 0.777	—	0.788,0.775
		NN	88.6	—	—	—
[17]	PIDD	RF	88.31	0.88	0.86	—	0.87
[2]	Private	LR	96.02	0.887	0.857	—	0.871
Proposed method	BRFSS	KNN	98.36	0.98	0.98	0.98	0.98