Optimal length of R–R interval segment window for Lorenz plot detection of paroxysmal atrial fibrillation by machine learning

Table 5 Effects of premature beats on classification accuracy between AF and non-AF LP images

Segment window length, beat	Ratio of premature beats in LP image
Segment window length, beat	0%	0–5%	5–10%	10–15%	15–20%	20–25%	25–30%	≥ 30%
10	0.988	–	0.869	–	0.674	–	0.495	0.388
20	0.994	0.955	0.926	0.886	0.809	0.725	0.704	0.694
50	0.996	0.972	0.936	0.900	0.861	0.791	0.748	0.822
100	0.997	0.975	0.936	0.926	0.896	0.840	0.809	0.836
200	0.997	0.974	0.919	0.919	0.806	0.688	0.773	0.768
500	0.997	0.969	0.912	0.819	0.668	0.485	0.648	0.703

Data are classification accuracy between AF and non-AF images annotated with the strict criteria

ISSN: 1475-925X