Stroke Prediction Using Smote-Tomek and Neural Network

• Published in: 2021 12th International Conference on Computing Communication and Networking Technologies (ICCCNT) pp. 1 - 5
• Main Authors: Rana, Chirag, Chitre, Nikita, Poyekar, Bhargavi, Bide, Pramod
• Format: Conference Proceeding
• Language: English
• Published: IEEE 06.07.2021
• Subjects: Class Imbalance; Computational modeling; Deep learning; Deep Neural Network; Machine Learning; Machine learning algorithms; Pandemics; Prediction algorithms; Predictive models; SMOTE-TOMEK Sampling method; Stroke; Stroke (medical condition)
• DOI: 10.1109/ICCCNT51525.2021.9579763

Stroke is a medical condition that occurs due to inadequate blood supply to the brain causing the death of brain cells. If a stroke is not diagnosed correctly can lead to brain injury, paralysis or even death. As the symptoms of Stroke are very instantaneous and can be triggered by unforeseen conditions also, it makes prevention of this situation very difficult to predict. The pandemic has made this worse as depression is the next stage of it which can be seen by the suffering of the people all across the globe. In our paper, we have taken a step to highlight this important topic of discussion by using advanced Machine learning and Deep Learning techniques to predict any possibility of stroke. We have proposed our final model using Artificial Neural Network which gives the best roc score of 0.84 and given a comparative analysis of how well do other Machine Learning Algorithms like ensemble-based, tree-based and Naive Bayes-based Algorithms perform in predicting Stroke.