Enhancing Heart Disease Prediction Based on B-Spline Curves and Machine Learning Algorithms

Abstract

Predicting heart attacks, is crucial as it can save lives and reduce the personal and societal impact of cardiovascular diseases. Early detection allows for timely intervention, enabling individuals to make lifestyle changes and medical professionals to implement preventive measures. In this presented work, we propose a new model for detecting and classifying heart diseases by integrating mathematical equations related to B-spline curves with artificial intelligence methods represented by machine learning. Since this method was proposed to improve the quality of medical datasets and to treat random and outlier values, the proposed approach consists of two stages for processing medical data: In the first stage, machine learning algorithms were trained on the dataset using classical pre-processing methods with (logistic regression) and (decision tree) algorithms, and the second stage involves applying the proposed method of combining pre-processing techniques which include b-spline curves and the Random Forest algorithm on the same dataset. The proposed approach achieved excellent results through the Random Forest model, with results indicating a detection accuracy of 99.19%. This proposed method of combined strategy provides valuable insights for early intervention. This marks a significant advancement in blending mathematical interpolation with machine learning, promising enhanced prediction accuracy and practical utility in healthcare analytics.