Exploring Heart Axis Classification on PTB-XL ECG Dataset Through Advanced Machine Learning

The aim of this study is to carry out a comparative study of the accuracy performance of the Random Forest and Logistic Regression model in the task of classifying the Heart Axis, which is a significant feature in cardiovascular health determination towards accuracy and facilitate observance. There were two groups in the experiment under observation. Group 1 used a Logistic Regression model, trained and validated on PTB-XL ECG dataset with 18 samples, features like heart rate, R-R intervals, and QRS durations were used for model deployment. Group 2 used a Random Forest model however, trained with the same features. They were also tested against baseline performance metrics like recall, accuracy, precision, and F1-score and Statistical significance level with 95% confidence. Logistic Regression performed at 88%. Random Forest classifier performed very high accuracy at 96%, much better than Logistic Regression and also performed better in recall and precision, enhances high ability of detection of heart axis abnormality. It was also enhanced as it could handle various ECG data, as it could handle various patient conditions and numerous recording differences. A Random Forest model can be compared to a Logistic Regression in heart axis classification with a greater accuracy, recall, and precision. The better performance reflects the Random Forest model as a prime candidate for real-time automatic heart axis classification systems. The research also reflects the potential of machine learning algorithms to make cardiovascular health analysis appear magical.