Predictive Analysis of Hypertensive Heart Disease Using a Machine Learning Approach

Hypertensive Heart Disease has become one of the contributing factors of heart failure disease, responsible for approximately 8 million deaths every year. and can lead to severe complications, including stroke, coronary heart disease, diabetes, kidney failure and blindness. This study aims to build a Machine Learning model by using CRISP-DM framework to predict suspected Hypertensive Heart Disease, and to explore the relationship of Hypertensive Heart Disease and blood cholesterol levels. Data were collected between January 2021 to June 2023 from a private hospital in Medan. Total data were 32,905 patients with hypertension, and 271 patients were diagnosed with Hypertensive Heart Disease. The data were processed using the RapidMiner. In this study, the algorithm with the highest accuracy and the lowest RMSE was selected for performance evaluation, and for the results showed that K-NN with Accuracy 89.65% and RMSE 0.280, had better performance than the Random Forest, Naive Bayes and Decision Tree algorithms. In summary, the implementation of Machine Learning can capture early detection and may help healthcare professionals to take necessary treatment regarding Hypertensive Heart Disease to improve patient prognosis and reduce the experiencing symptoms. Moreover, the hospitals and healthcare institutions can optimize their resource allocation for appropriate treatment and actions, leading to cost efficiency.