Year: 2025
Type: conference-paper
Abstract: Coronary heart disease(CHD) is one of the most common cardiovascular diseases. Studying coronary heart disease can better understand its pathogenesis, identify and control risk factors, thereby facilitating prevention and treatment to enhance patients’ quality of life. This study focuses on the data of the Behavioral Risk Factor Surveillance System Codebook Report in the United States in 2015 and systematically explores the impact of various factors on the occurrence of CHD. Firstly, the original dataset is preprocessed by feature engineering, and 21 related feature variables are selected and cleaned. To solve the problem of imbalanced target variable categories, four sampling methods are utilized to effectively improve the recognition ability of the model for the patients, including Synthetic Minority Oversampling Technique (SMOTE), Edited Nearest Neighbors (ENN), combining SMOTE and ENN (SMOTEENN), and MixUp. In the modeling phase, four machine learning models, random forest, logistic regression, gradient boosting, and eXtreme Gradient Boosting (XGBoost), are applied for training and prediction. By comparing the results of different models and sampling strategies through the model evaluation indicators, the impact of each feature variable on CHD is demonstrated. The study also introduces SHapley Additive exPlanations (SHAP) analysis to deeply analyze the decision-making process of the model and clarify the specific role of each feature variable in disease prediction.
Author: Yuling Qiu
Subfield: Health Information Management
Field: Health Professions
Domain: Health Sciences
Open Access status: gold
OpenAlex ID: W4417417814