Interactive cardiovascular disease prediction system using learning techniques: Insights from extensive experiments

Abstract

In Today's medical field, cardiovascular disease prediction is a significant challenge due to the influence of multiple variables affecting the circulatory system, such as hypertension, hyperlipidemia, and irregular pulse rates. Accurately classifying cardiac diseases proves to be a complex task. Consequently, the deep and machine learning techniques hold substantial potential for facilitating early identification. In this research paper, we explore the effectiveness of various models of machine learning, ensemble machine learning, and deep learning for predicting heart disease. These models undergo comprehensive experiments and cross-validation to evaluate their performance. To prepare the dataset, we apply standard scaling to numerical features, aligning them on a similar scale and enhancing the performance of specific learning algorithms. Our results demonstrate that deep learning models achieve high accuracy and robustness in predicting cardiovascular disease risk, with the InceptionNet model achieving an impressive 98.89 % accuracy. Additionally, ensemble learning models also show promise, with the Random Forest model delivering competitive accuracy, effectively capturing attributes and temporal dependencies within cardiovascular disease data. The findings of this study underscore the possibilities of deep learning and ensemble machine learning approaches in accurately predicting heart disease risk. Ultimately, this contributes to improved patient care and reduced mortality rates amidst the rising prevalence of heart-related conditions.