Predicting Cardiovascular Disease Risk in Tobacco Users Using Machine Learning Algorithms.

Area of Science:

• Cardiology
• Bioinformatics
• Public Health

Background:

• Cardiovascular diseases (CVDs) are a major global health concern.
• Tobacco use is a significant risk factor for CVDs.
• Existing predictive models often lack comprehensive risk factor integration, especially for high-risk populations like tobacco users.

Purpose of the Study:

• To identify additional phenotype factors that predict CVD risk in tobacco users.
• To explore the predictive power of various machine learning algorithms for CVD risk.
• To understand the interplay of risk factors contributing to cardiovascular events in tobacco users.

Main Methods:

• Utilized phenotype data from over 15,000 tobacco users in the UK Biobank.
• Applied multiple machine learning algorithms: Decision Trees (DT), Gradient Boosting (GB), Logistic Regression (LR), Random Forest (RF), and Support Vector Classification (SVC).
• Analyzed individual phenotype feature importance for CVD risk prediction.

Main Results:

• Identified specific phenotype factors with predictive power for CVD in tobacco users.
• Machine learning models demonstrated effectiveness in predicting CVD risk.
• Provided insights into the importance and interactions of various risk factors within this population.

Conclusions:

• Machine learning can effectively predict cardiovascular disease risk in tobacco users by integrating diverse phenotype data.
• Understanding the interplay of risk factors is crucial for targeted interventions in high-risk populations.
• This research contributes to improved CVD risk assessment and prevention strategies for tobacco users.