An arrhythmia classification algorithm using a dedicated wavelet adapted to different subjects

Table 5 The classification results of the proposed algorithm by heartbeat type and a comparison with other studies

Proposed algorithm's							Chazal *et al.*(2006) [7]
Real	Results					Total	Real	Results					Total
	N	S	V	F	Q			N	S	V	F	Q
N	44082	115	61	0	0	44258	N	32432	1360	66	483	53	34394
S	302	1448	87	0	0	1837	S	136	1264	31	10	0	1441
V	284	6	2924	7	0	3221	V	45	65	2433	32	4	2579
F	321	0	49	18	0	388	F	24	3	49	216	0	292
Q	5	0	2	0	0	7	Q	4	0	1	0	0	5
Total	44994	1569	3123	25	0	49711	Total	32641	2692	2580	741	57	38711
performance (%): Proposed algorithm							performance (%): Chazal et al. (2006)
	N	S	V	F	Q	Avg		N	S	V	F	Q	Avg
AC	97.81	98.97	99.00	99.24	99.99	97.94	AC	94.39	95.85	99.24	98.45	99.84	94.8
SF	83.28	99.75	99.57	99.99	100	85.07	SF	95.16	96.17	99.59	98.63	99.85	95.52
ST	99.60	78.82	90.78	4.64	0.00	97.51	ST	94.30	87.72	94.34	73.97	0.00	93.89
+P	97.97	92.29	93.63	72.00	0.00	97.26	+P	99.36	46.95	94.30	29.15	0.00	96.53
Chazal et al. (2004) [8]							Ince et al. (2009) [6]
Real	Results					Total	Real	Results					Total
	N	S	V	F	Q			N	S	V	F	Q
N	38444	1904	303	3509	98	44258	N	73019	991	513	98	29	74650
S	173	1395	252	16	1	1837	S	686	1568	205	5	6	2470
V	117	321	2504	176	103	3221	V	462	333	4993	79	32	5899
F	33	1	7	347	0	388	F	168	28	48	379	2	625
Q	4	0	3	0	0	7	Q	8	1	3	1	1	14
Total	38771	3621	3069	4048	202	49711	Total	74343	2921	5762	562	70	83658
performance (%): Chazal et al. (2004)							performance (%): Ince et al. (2009)
	N	S	V	F	Q	Avg		N	S	V	F	Q	Avg
AC	87.65	94.63	97.42	92.47	99.58	88.58	AC	96.47	97.30	98.00	99.49	99.90	96.62
SF	94.00	95.35	98.78	92.50	99.59	94.35	SF	85.30	98.33	99.01	99.78	99.92	86.76
ST	86.86	75.94	77.74	89.43	0.00	85.88	ST	97.82	63.48	84.64	60.64	7.14	95.58
+P	99.16	38.53	81.59	8.57	0.00	95.06	+P	98.22	53.68	86.65	67.44	0.00	95.84

AC, SF, ST and +P means accuracy, specificity, sensitivity and positive predictive value respectively.

ISSN: 1475-925X