Author(s): Imran Khan, Hitesh Motwani, Sahil Dhamecha, Sumit Jaisinghani, Devansh Motwani, Anish Mankani and Juhi Sawlani

Abstract: There are several diseases worldwide, many of which have well-established treatments. However, it is difficult to determine whether a person is experiencing a heart attack in the absence of a doctor. Our model can predict whether a person is experiencing a heart attack, even in the absence of a doctor. For example, Apple Watches can assist in the early identification of heart attack symptoms. We used large heart disease dataset from Kaggle.com and then performed Data preprocessing in it and we used various Classifiers for prediction. We used multiple classifiers among many others, including Random Forest, Decision Tree, and Logistic Regression The accuracy can be measured by using metrics such as F1 Score, Precision. Confusion matrix, Classification report. The model achieved accuracy of approximately 85% to 90% which is prominent given the dataset contains more than ten predictive features. Between the classifiers used, Random Forest achieved the highest accuracy. Logistic Regression and KNN obtained 86% and 71% accuracy, correspondingly. We also used other classifiers, such as SVM and K-means, but the highest accuracy was observed in Random Forest, Logistic Regression, and KNN. Accuracy can depend on multiple factors like independent variable, data preprocessing techniques, feature selection etc. So, it is basically depended on data and we have found best classifier for it. Our study's utilization of contemporary machine learning methodologies and active visualization tools, such Matplotlib and Seaborn, to enhance interpretability is a significant strength. To improve the display of the results, we utilized Matplotlib and Seaborn for visualization. In this outcome shows that the Random Forest and decision tree classifiers has maximum presentation for use to predict heart disease prediction. Heart disease prediction is crucial, especially in rural areas where access to medical professionals is limited. Our model can help in early detection, allowing individuals to seek medical attention sharp. At that time our model helps them to recognize whether a person has a heart attack or not. In this work, the predictive power of machine learning for heart disease is demonstrated using medical and demographic data. The results show that models such as decision trees and random forests may provide very accurate predictions, which could aid physicians in making earlier diagnoses.

DOI: 10.33545/27076571.2025.v6.i1a.136

Pages: 66-70 | Views: 520 | Downloads: 202

Download Full Article: Click Here