Figure 6
From: Design and evaluation of a knowledge-based ECG noise filtering framework
Feature response between dataset-specific sampling frequency (DSSF) and unified sampling frequency 500 Hz. (a) DSSF; (b) unified sampling frequency (500Hz) across the ECG datasets. From the kernel density estimate plot, it is clear that DSSF created three scatter points for noise-free ECG, while sampling frequency at 500 Hz shows only one point for noise-free ECG. So, it is easy for the model to separate two classes at 500Hz compared to the DSSF.
