Heart Disease Prediction Using Machine Learning on Heroku Using FLASK

Tham, Justin Jia Wei (2023) Heart Disease Prediction Using Machine Learning on Heroku Using FLASK. Final Year Project (Bachelor), Tunku Abdul Rahman University of Management and Technology.

Text RDS_Justin Tham Jia Wei_Fulltext.pdf Restricted to Registered users only Download (6MB)

Abstract

Heart Disease Prediction System is to use the modern approach to do the prediction because of the big amount of data and medical parameters. It can allow the users to reduce the risk of heart disease of a person due to early knowledge of the heart disease result. There are several functions in this Heart Disease Prediction System. Registration, prediction, data visualization are the main features in this system. Total of 7 Machine Learning algorithms have been used such as Random Forest, Gradient Boosting, Decision Tree, Support Vector Machine, GaussianNB, K-Nearest Neighbors, and Logistic Regression. Machine Learning, Heroku, and FLASK have been done in this system. Research has been done when doing the prediction using Machine Learning. When deploying the web application on Heroku, documentation also needs to be read since it is a new API tool during my degree year. After evaluating all the performance of algorithms using Machine Learning. Support Vector Machine with the Hyperparameter tuning of GridSearchCV has the best performance for this dataset to predict heart disease. Identifying the risk and the health condition of the patient is one of the strengths of the Heart Disease Prediction System. The second strength of the system is to avoid potential hospital admissions. Whereby, there are still weaknesses in this Heart Disease Prediction System. The dataset of this prediction is mainly from Cleveland. The result may not be accurate much since Cleveland has different habits with Malaysia. It will cause the prediction to not be accurate when Malaysia's citizens use the heart disease prediction system. Keywords: Data mining, Disease Diagnosis, Prediction, Machine Learning, FLASK, Heroku