ABSTRACT: The electrocardiogram (EKG) are indeed a crucial tools for detection cardiovascular issues, and our projects to digitize and analyze EKG papers records using machinery learn techniques sounds promisingly. By converts paper records into digital signal and applications vary techniques like feature extract, dimensions reducing, and classification algorithms, you are aims to automate the diagnoses processes, potential saving times and improving accuracies. our approach to splits the EKG report into lead, extractive waveforms (P, QRS, and T waving), and then converts them into a 1-D signals is logically. Using techniques such as smoothening, thresholds, and scales can help improving the qualities of the extracted signals, make them more suitable for analyzing. Apply dimension reductions techniques like Principle Components Analysis (PCA) is sensible for understands the data better and potential improving the efficiencies of the classifications processes. Employing multiple classifiers like k-nearest neighbor's (KNN), Logistic Regressions, Support Vector Machines (SVM), and an Voting Based Ensemble Classifier allows for comprehensive evaluations and comparisons of differs models. Assess the models based on metrics like accuracy, precision, recalling, f1-scores, and supporting is crucially for determined their effectiveness in diagnosed cardiac diseases. Ultimately, y'all final models aims to accurately diagnose conditions like Myocardial Infarctions, Abnormally Heartbeats, or determining if the patients is healthy based on the EKG reports. By translates the EKG findings into layman's terms, your systems could provides valuable insights to healthcares professionals and patients alike, aides in timely intervention and treatments.

Keywords: CNN, Logistic Regression, SVM, Streamlit, K-nearest neighbours (KNN)

| DOI: 10.17148/IARJSET.2024.11478