This project focuses on utilising machine learning to predict cardiovascular disease based on key risk factors. Cardiovascular disease is a leading global cause of preventable deaths, responsible for significant suffering and straining healthcare systems. It claims about 17.7 million lives annually, making up 44% of non-communicable disease fatalities.
We examined risk factors such as blood pressure, obesity, age, gender, diet, exercise, smoking, insurance, mental and physical health, alcohol use, sleep, and health check-ups. Our aim is to leverage machine learning techniques for accurate disease prediction, contributing to research and prevention efforts. Below is an organized breakdown of the project's key components and features:
Objective: The research aims to harness machine learning for predicting cardiovascular diseases based on numerous risk factors, addressing the high fatality rates associated with cardiovascular conditions.
Data Preprocessing: The paper details a rigorous preprocessing phase that includes resampling techniques to address dataset imbalances, standardization, and feature selection through Recursive Feature Elimination, setting a solid foundation for model training.
Machine Learning Models Implemented:
Decision Tree Classifier is used for its simplicity and interpretability. Logistic Regression is employed for its efficiency in binary classification tasks. Support Vector Machine (SVM) is chosen for its effectiveness in high-dimensional spaces.
Model Evaluation: The paper employs precision, recall, and F1-score to evaluate model performance, ensuring a comprehensive assessment beyond mere accuracy due to the imbalanced nature of the dataset.
Hyperparameter Tuning: Through Grid Search, the study fine-tunes model parameters, significantly impacting the models' predictive performances.