Machine Learning Approach for Estimation and Novel Design of Stroke Disease Predictions using Numerical and Categorical Features

Today, adequately trained machine learning algorithms can be significantly used in fields such as surveillance, medicine, and data management to identify and provide solutions to problems that do not have a solution answer current solutions are ineffective. A stroke is when blood arteries in the brain burst, harming the brain. It may also occur if the brain’s supply of nutrients and blood is interrupted. The severity of a stroke can be lessened by early recognition of numerous warning symptoms. This is a study using machine learning algorithms where children to adult age data have been taken in which we can extract data directly from their health report and after gathering all of the data, we run different algorithms models which will learn from this data and in future will show us how much is probability of you getting brain stroke. Five thousand people’s data were taken for processing, and we got that People with higher glucose levels are at increased risk of stroke, and high age Females are at risk of stroke. The machine learning algorithms we have considered are Random Forest Algorithm (RF), Decision Tree (DT), Support Vector Machine (SVM), and Logistic Regression (LR) Algorithm to train different models and compare the results for the best prediction model. Amongst all the algorithms, RF, LR, and SVM give us the best accuracy of 94.50% and 91.03% with the DT classifier. The Machine learning technique used in this article can assist medical professionals and patients with helpful analytical information and the probability of them getting a brain stroke in the future. Larger data sets and attribute selection strategies can enhance research with better accuracy, which can be achieved by taking data from the user’s input.”.