<![CDATA[Stroke is the second most deadly disease in the world and the third leading cause of disability. However, most deaths due to stroke can be prevented by recognizing the symptoms of stroke and taking preventive measures using information technology. Therefore, this research utilizes the role of information technology using a machine learning approach to predict stroke in a person using the K-Nearest Neighbor and Decision Tree classification methods. The two algorithms were compared to determine which algorithm was more effective in predicting stroke. Data analysis using the CRISP-DM approach was carried out using a dataset containing 5110 observations with 12 relevant attributes. Implementation of Exploratory Data Analysis (EDA) was also carried out for preprocessing, and oversampling techniques were applied to overcome the problem of unbalanced classes. The research results show that the predictive model with the highest level of accuracy was obtained at around 97.1845% using the K-Nearest Neighbor algorithm. This research makes a significant contribution to stroke prevention efforts through the use of information technology and machine learning algorithms for early identification of stroke risk.]]>
Copyrights © 2024
