Predicting hypertension, or high blood pressure, is crucial because it’s a risky condition that frequently doesn’t exhibit any obvious symptoms or indicators until major problems arise. By using predictive models for early identification, the risk of uncontrolled hypertension-related cardiovascular events, strokes, and organ damage can be decreased to take quick action. The study aims to predict the diagnosis of hypertension using three machine learning techniques: Logistic Regression, Gaussian Naive Bayes, and Support Vector Machine(SVM). With the focus on improving accuracy and reliability, the various hyperparameters of each individual algorithm are tuned using the cross validation and grid search approaches. After extensive testing and parameter modification, the SVM algorithm was found to be the most accurate predictor, which achieved a significant accuracy of ${9 7. 0 0 \%}$. The best hyper-parameters achieved for SVM are c value 4 and kernel as ‘rbf’. The models’ performance is evaluated using accuracy, specificity, sensitivity, F1-score, or AUC (Area Under the Curve). This study facilitates proactive healthcare measures by offering insightful information to determine the best algorithms for the early prediction of hypertension.
Analysis of Various Hyperparameters for Predicting Hypertension using Machine Learning Algorithms
2024-11-06
494802 byte
Conference paper
Electronic Resource
English
Computer Optimization of ANN Hyperparameters for Retrospective Information Processing
Springer Verlag | 2022
|Comparison of Machine Learning Algorithms for Predicting Lane Changing Intent
Springer Verlag | 2021
|Predicting Pavement Roughness Using Deep Learning Algorithms
Transportation Research Record | 2021
|American Institute of Physics | 2023
|