Patient-Specific Deep Architectural Model for ECG Classification

Table 5

Classification metrics compared to the state-of-the-art four-class heartbeat recognitions (percentage, %).

Methods	Accuracy	g-mean		N		S		V		F
		SE	PPV	SE	PPV	SE	PPV	SE	PPV	SE	PPV
Proposed (a)	97.5	85.9	84.4	99.0	98.4	71.4	94.4	93.3	93.3	82.7	58.5
Chazal and Reilly [11]	93.9	87.2	59.8	94.3	99.4	87.7	47.0	94.3	96.2	74.0	29.1
Jiang and Kong [8]	94.5	62.7	83.8	98.7	96.2	50.6	68.0	86.6	89.4	35.8	84.2
Ince et al. [7]	93.6	74.5	76.9	97.0	97.0	62.1	56.7	83.4	86.5	61.4	73.4
Proposed (b)	89.3	14.6	16.2	95.3	93.0	15.4	47.3	60.4	66.8	0.5	0.2
Mar et al. [12]	89.0	79.3	45.2	94.2	99.2	86.2	56.7	92.4	93.4	66.4	17.7
Alvarado et al. [10]	93.6	84.0	55.2	94.2	99.2	86.2	56.7	92.4	93.4	66.4	17.7
Ye et al. [9]	88.2	62.6	37.0	90.0	98.2	56.4	55.1	84.7	59.5	35.8	5.8
Zhang et al. [13]	88.3	86.7	46.2	88.9	99.0	79.1	36.0	85.5	92.8	93.8	13.7

Patient-specific method: require expert intervention.
(a) indicates the patient-specific heartbeat classification scenario. Classifiers were trained by using the first 300 beats of individual patient.
(b) indicates interpatient heartbeat classification scenario.