Application of machine learning algorithm on binary classification model for stroke treatment eligibility

Abstract

In Canada, stroke is the leading cause of adult disability and the third leading cause of death. Ischemic stroke is the most common type, making up approximately 85% of all stroke patients. Endovascular treatment (EVT) is effective for severe ischemic stroke patients. Unfortunately, EVT requires specialized equipment and personnel, which limits its availability. There are several clinical and imaging factors that are critical in determining eligibility for EVT. Furthermore, in stroke, minutes matter as the brain dies quickly after onset, making EVT treatment's effectiveness highly time dependent. For this reason, timely across to EVT is critical. This study is to create a binary classification model to predict the EVT eligibility of stroke patients and discover attributes of the patient information that help to make efficient decision on transfer EVT eligible patient. Following algorithms applied to dataset: Logistic Regression, Decision Tree, Random Forest, and Support Vector Machine.