Skip to main content

Table 2 Descriptive analysis of the features of the validation dataset

From: A filter approach for feature selection in classification: application to automatic atrial fibrillation detection in electrocardiogram recordings

Features

AF (N = 11)

NSR (N = 23)

p-value

\(m_0\)

\(0.769\ (0.248)\)

\(0.873\ (0.114)\)

0.028

\(m_1\)

\(-0.036\ (0.016)\)

\(-0.002\ (0.002)\)

0.000

\(m_2\)

\(0.109\ (0.061)\)

\(0.006\ (0.007)\)

0.000

\(m_3\)

\(-0.273\ (0.188)\)

\(-0.019\ (0.023)\)

0.000

\(m_4\)

\(0.656\ (0.536)\)

\(0.065\ (0.085)\)

0.000

\(m_5\)

\(-1.633\ (1.508)\)

\(-0.252\ (0.356)\)

0.000

\(m_6\)

\(4.331\ (4.453)\)

\(1.104\ (1.602)\)

0.002

\(m_7\)

\(-13.642\ (16.670)\)

\(-5.073\ (7.710)\)

0.015

\(m_8\)

\(53.254\ (91.939)\)

\(24.687\ (38.695)\)

0.071

\(m_9\)

\(-288.073\ (666.902)\)

\(-124.233\ (203.657)\)

0.243

\(m_{10}\)

\(1901.032 \ (5304.157)\)

\(648.314\ (1122.787)\)

0.690

\(sd_0\)

\(0.142\ (0.037)\)

\(0.047\ (0.019)\)

0.000

\(sd_1\)

\(0.270\ (0.069)\)

\(0.042\ (0.020)\)

0.000

\(sd_2\)

\(0.714\ (0.296)\)

\(0.097\ (0.057)\)

0.000

\(sd_3\)

\(2.163\ (1.235)\)

\(0.297\ (0.240)\)

0.000

\(sd_4\)

\(7.190\ (5.080)\)

\(1.077\ (1.071)\)

0.000

\(sd_5\)

\(25.866\ (21.270)\)

\(4.355\ (4.916)\)

0.000

\(sd_6\)

\(99.809\ (91.928)\)

\(18.842\ (23.025)\)

0.001

\(sd_7\)

\(411.791\ (416.661)\)

\(85.683\ (111.063)\)

0.005

\(sd_8\)

\(1822.127\ (2043.844)\)

\(404.282\ (549.422)\)

0.008

\(sd_9\)

\(8773.333\ (11{,}389.615)\)

\(1976.371\ (2815.619)\)

0.026

\(sd_{10}\)

\(46{,}563.114\ (73{,}918.526)\)

\(9916.219\ (14{,}751.188)\)

0.042

SDNN

\(141.717\ (37.248)\)

\(46.580\ (18.907)\)

0.000

SDANN

\(60.200\ (24.675)\)

\(35.914\ (14.735)\)

0.010

SDNNIDX

\(131.858\ (32.635)\)

\(25.081\ (12.269)\)

0.000

pNN50

\(74.878\ (6.375)\)

\(10.601\ (11.328)\)

0.000

SDSD

\(195.287\ (47.089)\)

\(35.023\ (18.044)\)

0.000

RMSSD

\(193.983\ (46.599)\)

\(34.779\ (17.902)\)

0.000

IRRR

\(185.401\ (54.637)\)

\(57.067\ (26.341)\)

0.000

MADRR

\(123.929\ (30.981)\)

\(17.391\ (11.339)\)

0.000

TINN

\(247.492\ (27.711)\)

\(114.821\ (33.433)\)

0.000

HRV.index

\(15.839\ (1.773)\)

\(7.349\ (2.140)\)

0.000

  1. For each patient, we computed the mean value of the features derived from each of his or her 1-min RR interval. Each line corresponds to the mean values and standard deviations of all patients in the AF and NSR groups. The p-value corresponds to a Student’s t-test in case of normal of normal distribution and to a Mann–Whitney U test in case of non-normal distribution