Predicting Heart Disease Risk Using Machine Learning and Regression Analysis

Chin, Jia Hao (2025) Predicting Heart Disease Risk Using Machine Learning and Regression Analysis. Final Year Project (Bachelor), Tunku Abdul Rahman University of Management and Technology.

Text RMM_Chin Jia Hao_Full Text.pdf Restricted to Registered users only Download (2MB)

Abstract

Heart disease remains a leading cause of mortality worldwide, highlighting the need for accurate predictive models for early detection and prevention. This study develops a machine learning-based predictive model using Logistic Regression, Support Vector Classification (SVC), and Neural Networks to assess heart disease risk based on patient health metrics. Chapter 1 introduces the research background, problem statement, objectives, scope, and significance. It emphasizes the role of machine learning in improving prediction accuracy compared to traditional diagnostic methods. Chapter 2 reviews existing literature on heart disease prediction, examining machine learning techniques, prior research findings, and potential advancements in predictive modelling. Chapter 3 details the research methodology, covering dataset selection, preprocessing, and model implementation. The dataset includes clinical features such as age, cholesterol levels, blood pressure, and other cardiovascular indicators. Preprocessing steps like normalization, feature selection, and standardization are applied to enhance model performance. Hyperparameter tuning and cross-validation optimize the models. Chapter 4 presents results and discussions, evaluating models using accuracy, precision, recall, and F1-score. Logistic Regression with retaining the features ‘fbs’ achieves the highest accuracy and optimal balance between precision and recall, proving most effective in predicting heart disease risk. The influence of feature selection and hyperparameter tuning on model performance is analysed. Chapter 5 concludes the study by summarizing key findings, addressing challenges like data imbalance and limited real-world variables, and suggesting future research directions. The study highlights machine learning’s potential in clinical decision-making, recommending the integration of additional features and deep learning techniques to further enhance prediction accuracy. Keywords: Heart Disease Prediction, Machine Learning, Logistic Regression, Support Vector Classification (SVC), Neural Networks, Medical Data Analysis, Risk Prediction.