Machine learning-based approaches for automated hypertension detection

Le résumé fourni par la source

Early detection and timely intervention of hypertension is critical to mitigate its adverse effects as it is a leading cause of heart disease. Traditionally, to detect hypertension depends upon various clinical factors such as measuring blood pressure, assessment of medical history, etc but they also have certain limitations such as human error, time consuming, as well as chances of missed diagnosis. Hence, using AI techniques allows identifying complex relationships within the data as well as surpassing the limitations of traditional diagnostic methods. The aim of the paper is to create a model which is capable of identifying individuals with hypertension and those without by using various AI learning models. With a dataset of 14 attributes, preprocessing identifies missing data while as visualization and correlation analysis has been used for detecting outliers and feature selection respectively. Further, seven different classifiers are applied such as gradient boosting, voting classifier, light gradient boosting machine, multilayer perceptron, Catboost, eXtreme gradient boosting classifier, and Adaboost. The performance of these classifiers has been evaluated it appears that CatBoost, XGBoost, and LGBM perform with a remarkable accuracy of 100% as well as outstanding recall, precision, along with the F1 score of 1.00. It demonstrates the capability of AI-driven models to revolutionize hypertension detection.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.