Machine Learning Models for Early Brain Stroke Prediction: A Performance Analogy
Somya Srivastav, Kalpna Guleria, Shagun Sharma
Abstract
When the supply of blood gets stopped to the brain, then the brain strokes happen. This results in damage to the nerve cells due to which cells do not get enough oxygen and nutrition they require. This can be a cause of death, major illness, or permanent brain damage. To avoid the challenges of brain stroke, it is required to predict it in the early stages. In this work, the dataset has been obtained from Kaggle which is further projected to the proposed prediction model. This study developed a prediction model that can aid with predicting whether a person might suffer a brain stroke or not. Classification accuracy, F1-measure, recall, precision, and root means square error (RMSE) analysis have been employed to evaluate the prediction performance of the proposed model. An analytical comparison has been also done using the outputs of multiple machine-learning algorithms implemented in the proposed work. Furthermore, various types of algorithms proposed in the work can effectively predict brain stroke in patients. Conclusively, when the accuracy comparison has been done the “Logistic regression” gets the best results at 95.02% in comparison to other machine-learning techniques.