Exploratory Data Analysis for Brain Stroke Prediction Using Multiple Regression Models

Stroke has been one of the leading causes of death. Recognizing its warning signs early can lessen its severity and maximize the effectiveness of its prevention. A highly effective data-driven predictive algorithm is needed for this, which can be done through various machine learning-based techniques. In this paper, Linear Support Vector Machines (linear-SVM), Radial Support Vector Machines (rbf-SVM), Logistic Regression, Decision Tree, Gaussian Naive Bayes, and Random Forests were used for stroke prediction. Among these models, linear-SVM, rbf-SVM, and Logistic Regression yielded the highest accuracy value of 95%. Cross-validation and boosting through ensembling were also conducted to further enhance the overall confidence of model prediction.