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

Table 6 The classification results of #232 with or without information of past heartbeats (Annotbf1, Annotbf2, Morphbf1 and Morphbf2)

With information of past heartbeats							Without information of past heartbeats
Real	Results					Total	Real	Results					Total
	N	S	V	F	Q			N	S	V	F	Q
N	397	1	0	0	0	398	N	398	0	0	0	0	398
S	1	1379	2	0	0	1382	S	864	319	199	0	0	1382
V	0	0	0	0	0	0	V	0	0	0	0	0	0
F	0	0	0	0	0	0	F	0	0	0	0	0	0
Q	0	0	0	0	0	0	Q	0	0	0	0	0	0
Total	398	1380	2	0	0	1780	Total	1262	319	199	0	0	1780
performance (%): Proposed algorithm							performance (%): Chazal et al. (2006)
	N	S	V	F	Q	Avg		N	S	V	F	Q	Avg
AC	99.89	99.78	99.89	100	100	99.80	AC	51.46	40.28	88.82	100	100	42.78
SF	99.93	99.75	99.89	100	100	99.79	SF	37.48	100	88.82	100	100	86.02
ST	99.75	99.78	-	-	-	-	ST	100	23.08	-	-	-	-
+P	99.75	99.93	0.00	-	0.00	-	+P	31.54	100	0.00	-	0.00	-

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

ISSN: 1475-925X