Open Access
American Research Journal of Cardiovascular Diseases
ISSN (Online): 2575-7601
DOI: 10.46568/arjcd
Cardiovascular Disease Predictive Modeling with Machine Learning Feature Importance
Bergen County Academies, Hackensack, NJ 07601.
Daniel Han, “Cardiovascular Disease Predictive Modeling with Machine Learning Feature Importance”,
American Research Journal of Cardiovascular Diseases, Vol 5, no. 1, 2024, pp. 01-06.
Abstract
Cardiovascular diseases are one of the leading factors of death around the world. To provide insights on the correlation
between general factors and the existence of cardiovascular diseases, a dataset supplied by Svetlana Ulianova in Kaggle
with 70,000 patient records with 11 features and a target was used to determine what attributes have the most influence
on cardiovascular conditions. The research results suggest that 1) Smoking, height, and gender did not have a significant
contribution to cardiovascular disease 2) Systolic and diastolic were shown to have a strong contribution to cardiovascular
disease 3) Random Forest model performance yielded the highest metrics compared to both Gassausian Naive Bayes and
the benchmark model Logistic Regression. This research can assist Doctors with determining patients who have a high
susceptibility to cardiovascular disease.