Hybrid Model based on Support Vector Machine and Principal Component Analysis Applied to Arterial Hypertension Detection

Abstract

This research aims to reduce the detection time of the risk of suffering from arterial hypertension by implementing a hybrid model based on the Support Vector Machine (SVM) and Principal Component Analysis (PCA) algorithms. The proposed hybrid model was implemented from the processing of a dataset made up of 70,000 records related to characteristics such as systolic blood pressure, diastolic blood pressure, cholesterol index, glucose index, smoking and sedentary lifestyle. The methodology for the implementation of the hybrid model consisted of the stages of data collection, data exploration, data pre-processing, selection of characteristics, and implementation of the model and the validation of results. As a result of the implementation of the model, a precision level of 72.18% was obtained in relation to the detection of the risk of suffering from arterial hypertension.