| dc.description.abstract | This thesis presents a comprehensive evaluation of machine learning algorithms for the
early detection and diagnosis of stroke and diabetes. The study utilizes popular algorithms such as
Logistic Regression, Decision Tree, Random Forest, K-Nearest Neighbors (KNN), and XGBoost
Classifier to analyze medical data and extract valuable insights. The XGBoost Classifier emerged
as the top-performing model, achieving an exceptional accuracy, precision, recall, and F1-score of
87.5%. Comparative analysis of the algorithms revealed that the Decision Tree, Random Forest,
and XGBoost classifiers consistently demonstrated high performance across all metrics. These
models exhibited remarkable discrimination abilities, with the XGBoost Classifier and Random
Forest achieving accuracy values of approximately 87.5% and 86.5% respectively. The Decision
Tree Classifier also delivered strong performance with an accuracy of 83%. The F1-score, which
considers both precision and recall, reflected the overall accuracy of the models, with the XGBoost
model slightly outperforming the Random Forest and Decision Tree models by 2% and 4.25%
respectively. These findings highlight the efficacy of the XGBoost Classifier and therefore will be
used in this project as a model for prediction, along with the Random Forest and Decision Tree
models, in accurately diagnosing stroke and diabetes. The results of this study contribute to the
development of AI-powered diagnostic systems for improved healthcare decision-making. | en_US |
