Stroke Prediction using Machine Learning

In this work, we aimed to predict the incidence of strokes using machine learning approaches. The dataset includes demographic and health-related variables such as age, gender, heart disease, hypertension, and smoking status. After pre- processing the data, which included encoding categorical variables and handling missing values, we trained several classification techniques, including Random Forest Classifier, AdaBoost Classifier, and Gradient Boosting Classifier. To evaluate the models' performance, we employed metrics such as F1-score, recall, accuracy, and precision. The Random Forest Classifier achieved the highest accuracy of 94.72% on the test set. In order to solve the issue of class imbalance, we also employed techniques like Random Over Sampler and SMOTE (Synthetic Minority Over-Sampling Technique), which improved the models' capacity to predict the recurrence of strokes. All things considered, our findings suggest that machine learning algorithms, with the Random Forest Classifier showing promising accuracy results, may be able to predict the incidence of strokes based on demographic and health-related data.