Stroke Prediction Using Machine Learning Classification Methods

Based on machine learning, this paper aims to build a supervised model that can predict the presence of a stroke in the near future based on certain factors using different machine learning classification methods. The predictions resulting from this model can save many lives or give people hints on how they can protect themselves from the risk. The models obtained from this research are just a tool that doctors can use; thus, it does not take the role of doctors. The model was trained on a dataset that contains the factors or features that affect stroke disease. The correlation values were calculated to know how much a particular feature affects the target feature (having a stroke) or if other features are affected by it. After all, the model was tested on a set of samples to measure the accuracy of the trained model. Finally, multiple models were produced using different algorithms (classifiers), but the model that produced the best accuracy, precision, recall, and F1-score of 94%-95% is based on the Random Forest classifier.