Performance Analysis of Machine Learning Approaches for Stroke Prediction in Healthcare

A stroke is a situation that happens due to insufficient blood supply to the brain. Nowadays, there is an increased number of deaths which can be caused due to many reasons including stroke as one of the major reasons. Other impacts of stroke are facial paralysis and one side of body paralysis. This is a big concern as it could lead to an increased death rate. But, due to the advancements in the field of technology, it is easy to make predictions about a stroke. Various machine-learning techniques are used for this purpose. In this article, four major techniques are used Naive Bayes, KStar, multilayer perceptron, and random forest. To apply these classifiers the dataset is taken from Kaggle and by using the WEKA tool the performance is evaluated in the form of accuracy, recall, precision, F-measure, and execution time. In the proposed work, it is found that random forest works best with the highest accuracy of 95.02%. Further, the performance of the proposed model has also been compared with the existing works and it has been found that the proposed model exhibits better performance. Additionally, the work may contribute to the healthcare sector to provide assistance to clinicians for the early prediction of strokes.