Improving Accuracy in Ecg Signal for Heart Malfunctioning Using Machine Learning Using CNN Autoencoder Compared to CNN1d

The principal objective of this work is to increase the accuracy of the ECG signal for cardiac dysfunction. The Kaggle dataset served as the study's main data source. This study used two different groups, Group I and Group 2, each with 20 samples. Group 2 used the CNN1d, and Group I used the CNN Autoencoder. Forty people made up the study's entire sample size. Python was used for the implementation of the sample size computations for statistical analysis and the ensuing performance comparison. Clincalc.com was used to do the statistical analysis, using the following settings: alpha (a) at 0.05, beta (13) at 0.2, and statistical power (G -power) at 85 %.. The analysis primarily focused on comparing the performance of the CNN Autoencoder and Algorithm using accuracy value as the key evaluation metric. In terms of accuracy, CNN Autoencoder (97.002 %) outperforms CNN1d (85.179 %), with a two-tailed, <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{p}_{\boldsymbol{i}} 0.05$</tex> significance value of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{i}_{\mathbf{i}}. 001$</tex>. In summary, the accuracy of CNN Autoencoder outperforms CNN1d accuracy.