Project Report on Heart Failure Prediction System Using Machine Learning

Abstract

The Framework which we have used to develop a web application for the heart failure prediction system will be discussed in this report. In this users can check their age and heart rate from any fitness app and then can enter the details. There are many algorithms are there which we can choose but here we have chosen ensemble learning and machine learning by which the users can identify their heart health. If we talk about the project’s main goal, we have created a web application that will be very useful for users to check their cardiovascular health. It contains all the steps like data gathering, preprocessing, model creation, and creation of web application. The accuracy increases with ensemble learning and we can say that the heart failure system has all the possibility to be a very helpful web application for people to manage their cardiovascular health and take early preventions. The main goal of the project is to develop a web application in which the data like age and heart rate need to be entered and then it finally analyses it predicts the results. This actually is very useful for people to know the early detection and early prevention and to avoid heart failure or heart attack .because we are using ensemble learning so which increases the accuracy of the tool. So, there are some basic steps that we have used to develop this tool like data gathering, data pre preprocessing, creation of the model, and development of web application. In this, we gather data from UCI and many fitness tools to get the relevant data. So in this preprocessed data ensures the data consistency and quality of the data then on this preprocessed data the machine learning or ensemble learning models like SVM, logistic Regression, and Decision trees are made to be trained. The data which we have collected shows that the most number of ill people are male . We can draw many conclusions from the heart attack prediction system. First of all, we can say that compared to many single models, ensemble learning or machine learning approach massively increase the prediction accuracy. Just because we have used the SVM, decision tree, Logistic regression, KNN, and Naïve Bayes it boosts the prediction accuracy using ensemble learning techniques like soft voting and hard voting. Second we can conclude from the data we have gathered has most number of ill patients are men. and it depends on many factors like age, resting, cholesterol, blood pressure. The heart failure prediction system is a very promising approach that helps people to identify the likelihood of any issue regarding their cardiovascular health. Which helps them in early detection and early prevention of any disease regarding the heart. Real-time forecasts or insights are to be done through the accessible web application. In the coming future more study be done to enhance the models and add many more important features for improved accuracy. The usability and efficiency of the system may also be acquired through the user input in the real-time environment. To conclude the above statements, we can say that flask-based heart failure system deliver the customers with an accessible tool for identifying the risk of heart failure or heart attack. So in study finds out that there are many factors that can affect heart attack like age, blood pressure, cholesterol, and many more. This system has the potential to support taking charge of the heart health system and give users the necessary information that they need to take some actions regarding their cardiovascular health.