Predictive modeling of cardiovascular disease risk through lifestyle analysis using various machine learning algorithms

Cardiovascular disease (CVD) is a major contributor of global death, with statistics about 31% of annual deaths. The objective of this study is to build a predictive model for cardiovascular disease (CVD) risk by examining lifestyle factors, using data from the Behavioral Risk Factor Surveillance System (BRFSS) for the year 2021. The main objectives consist of identifying key lifestyle factors that impact the risk of cardiovascular disease (CVD), creating a highly accurate predictive model, and evaluating its performance using various metrics. Various machine learning algorithms were utilized to construct predictive models, including logistic regression, decision trees, random forest, gradient boosting, K-Nearest Neighbor (KNN), LightGBM, extra trees, AdaBoost, XGBoost, ridge classifier, Quadratic Discriminant Analysis (QDA), Feedforward Neural Network (FNN), neural network, and Long Short-Term Memory (LSTM). The analysis revealed that variables such as age, arthritis, smoking history, and other lifestyle factors have a huge impact on the risk of CVD. Out of all the models tested, Gradient Boosting and LightGBM classifiers emerged as the most effective, demonstrating the best balance between precision and recall for identifying CVD cases, with an accuracy rate of around 92%. This study highlights the potential of lifestyle modifications in reducing the risk of CVD. The advanced predictive models serve as valuable tools for early detection and prevention, offering practical insights for healthcare practitioners and policymakers.