Figure 4
Clustering of two-dimensional SHAP values for ECGs in the test dataset by which the model predicted reduced LVEF. The 192 ECGs by which the model predicted reduced LVEF, excluding those with a paced rhythm, were clustered by adapting a variational Bayesian Gaussian mixture model to the two-dimensional SHAP values. The clusters are classified from 1 to 6, and the data points belonging to each cluster are coloured red, green, blue, orange, pink, or brown. UMAP, uniform manifold approximation and projection.

Clustering of two-dimensional SHAP values for ECGs in the test dataset by which the model predicted reduced LVEF. The 192 ECGs by which the model predicted reduced LVEF, excluding those with a paced rhythm, were clustered by adapting a variational Bayesian Gaussian mixture model to the two-dimensional SHAP values. The clusters are classified from 1 to 6, and the data points belonging to each cluster are coloured red, green, blue, orange, pink, or brown. UMAP, uniform manifold approximation and projection.

Close
This Feature Is Available To Subscribers Only

Sign In or Create an Account

Close

This PDF is available to Subscribers Only

View Article Abstract & Purchase Options

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Close