Table 1 Dataset characteristics

From: Interpretable arrhythmia detection in ECG scans using deep learning ensembles: a genetic programming approach

Dataset

Guangzhou Heart Study Dataset

Liverpool Heart And Chest Hospital Study Dataset

Dataset acronym

Dataset G

Dataset L

Age of patients

71.4 ± 6.3 years

60.5 ± 10.71 years

Number of patients

1172(66% women)

909(32.9% women)

Number of ECG scans

1172

1807

Patients with arrhythmia [size/%]

241/20.5%

270/29.7%

Patients without arrhythmia [size/%]

931/79.5%

639/63.3%

Training (T) Size/% of arrhythmia

828/20.2%

1355/26.3%

Validation (V) Size/% of arrhythmia

172/20.3%

265/26.0%

Testing (Ѱ) Size/% of arrhythmia

172/22.1%

187/44.4%

Mean for channels R/G/B

0.94/0.91/0.91

0.82/0.75/0.76

Standard deviation for channels R/G/B

0.17/0.17/0.17

0.16/0.17/0.17

  1. Regarding the lifestyle characteristics in dataset L, 81% of patients (N = 733) reported current alcohol consumption. The smoking status was as follows: never-smokers (N = 506, 56%), current smokers (N = 66, 7%), and former smokers (N = 341, 37%).