Hypertension Prediction Using Stacked Ensemble Model from Imbalanced Clinical Data

• Published in: 2024 International Conference on Advances in Computing, Communication, Electrical, and Smart Systems (iCACCESS) pp. 1 - 5
• Main Authors: Ullah, Shah Muhammad Azmat, Hossain, A. B. M. Aowlad
• Format: Conference Proceeding
• Language: English
• Published: IEEE 08.03.2024
• Subjects: Computational modeling; Data models; Hypertension; hypertension prediction; imbalanced data; Performance evaluation; Predictive models; Refining; SMOTETomek; stacked ensemble learning; Support vector machines
• DOI: 10.1109/iCACCESS61735.2024.10499627

Early disease prediction is vital for improving healthcare quality and preventing patients from developing critical health issues. This research introduces a Hypertension Prediction Model (HPM) that uses individual clinical data to predict the hypertension at an early stage using a stacked ensemble machine learning technique. To resolve data distribution imbalances, the proposed framework employs the Synthetic Minority Oversampling Technique Tomek Link (SMOTETomek). The popular classification models K-Nearest Neighbour (KNN) and Random Forest (RF) are used as a base level classifier and Support Vector Machine (SVM) is used as a meta level classifier in stacked ensemble model. Three datasets of both male and female subjects and their combination were used to train and evaluate the proposed model with an emphasis to enhance the generalization capability of the classifier. Various performance evaluation matrices are used to assess and analyze the performance of the classifier under different dataset cases. The obtained results show that the proposed HPM achieves the superior accuracy when compared to alternative models and past research investigations.