Table 2 Comparison of AUC results achieved for different models, according to different scenarios.

From: Analyzing artificial intelligence systems for the prediction of atrial fibrillation from sinus-rhythm ECGs including demographics and feature visualization

Model

Random sampling

Balancing distributions

Young SR data

Resnet

0.87 (0.81–0.93)

0.77 (0.70–0.84)

0.92 (0.87–0.97)

Resnet with age/sex

0.89 (0.83–0.95)

0.79 (0.72–0.86)

0.98 (0.95–1.00)

MLR

0.86 (0.80–0.93)

0.71 (0.64–0.78)

0.97 (0.94–1.00)

  1. In each scenario (random sampling, balancing distribution, young SR data) the same ECG samples have been used to train and evaluate the three different AI models. The only exception refers to MLR, where ECG samples with missing discrete values have been discarded from the analysis. AUC results are computed according to the aggregation rule by average. 95% confidence intervals are reported in brackets.
  2. ECG Electrocardiogram, MLR Multivariate Logistic Regression.
Back to article page