Key result
Logistic regression with feature selection classifies cardiovascular disease with ~87% accuracy.
Why the study?
Cardiovascular disease accounts for nearly 17 million deaths worldwide, and early diagnosis is crucial for providing timely treatment and reducing mortality. This study evaluated a Logistic Regression model with correlation-based feature selection to classify the presence or absence of cardiovascular disease.
Does a Logistic Regression machine learning model with correlation-based feature selection accurately classify the presence or absence of cardiovascular disease in the UCI heart disease dataset?
Does a Logistic Regression machine learning model with correlation-based feature selection accurately classify the presence or absence of cardiovascular disease in the UCI heart disease dataset?
A Logistic Regression classifier utilizing correlation-based feature selection achieved 87.10% accuracy in predicting cardiovascular disease on the UCI dataset when using a 90:10 training-to-testing split.
Cardiovascular disease is one of the most life-threatening diseases and accounts for nearly 17 million deaths worldwide. Early diagnosis plays a crucial role in providing timely treatment and reducing mortality. Several machine learning and deep learning techniques are available to classify the presence or absence of cardiovascular disease. In this study, a Logistic Regression (LR) model is applied to the UCI dataset for heart disease classification. To enhance the model performance, data preprocessing steps such as dataset cleaning, missing value handling, and feature selection were carried out. Feature selection was performed based on the correlation between each feature and the target variable, and highly positively correlated features were selected. The dataset was then divided into training and testing sets using different split ratios: 90:10, 80:20, 70:30, 50:50, and 40:60. Among these, the 90:10 split ratio achieved the best performance. The proposed Logistic Regression model attained an accuracy of 87.10%.
No takes yet. Share an insight, caveat, or question.
P.Mangaiyarkkarasi et al. (2026) studied Cardiovascular disease (n=303). Logistic Regression classifier (90:10 training-to-testing split with correlation-based feature selection) vs. Alternative training-to-testing split ratios (80:20, 70:30, 60:40, 50:50) was evaluated on Classification accuracy for presence of cardiovascular disease. The Logistic Regression model with correlation-based feature selection achieved 87.10% accuracy (precision 0.857, recall 0.857) for cardiovascular disease classification using a 90:10 split.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: