Early Brain Stroke Detection Using Artificial Intelligence
No Thumbnail Available
Date
2024-04-30
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Due to the current environmental conditions and human lifestyle choices, people are impacted by
a wide range of diseases nowadays. If such diseases are to be prevented from reaching their
terminal phases, early detection and prediction are essential. An enormous financial burden is
placed on individuals who suffer from stroke, a cerebrovascular disorder that is one of the main
causes of mortality. One of the main risk factors for stroke is health-related behavior, which is
gaining importance as a prevention strategy. Many machine learning algorithms that incorporate
lifestyle characteristics as predictors to autonomously diagnose stroke have been used to predict
the risk of stroke.
This work uses five supervised machine learning classifiers to predict strokes: K-Nearest Neighbor
Algorithm, Decision Tree, Random Forest, Support Vector Machine, and Naïve Bayes. The
aforementioned classifiers are trained on the dataset, which consists of 5110 items with 10
attributes, and their performance is assessed using the confusion matrix. The dataset is
preprocessed to make it acceptable for prediction. The RF method surpassed all other algorithms
in the employed dataset for predicting strokes based on many physiological characteristics, with
an accuracy of 95.8%. In contrast to an individual's medical history and level of physical activity,
machine learning algorithms may be more useful for the clinical estimation of stroke.
Stroke patients need continuous critical care in addition to all of these diagnoses, which can be
provided by an interdisciplinary team.
Keyword: Artificial Intelligence, Confusion Matrix, Random Forest Classifier, Stroke