An Adaptive Heart Disease Behavior-Based Prediction System

Abstract

Heart disease prediction is a complex process that is influenced by several factors, including the combination of attributes leading to the possibility of heart disease and availability of these attributes in the database, an accurate selection of these attributes and determining the priority and impact of each of them on the prediction model, and finally selecting the appropriate classification technique to build the model. Most of the previous studies have used some heart disease symptoms as major risk factors to build a heart disease prediction system leading to inaccurate prediction results. The main objective of this study is to build an Adaptive Heart Disease Behavior-Based Prediction System (AHDBP) based on risk factors and behaviors that may lead to heart disease. Different classification algorithms will be deployed to get the most accurate results. 18 attributes were used to build the prediction system. The accuracy of the classification techniques was as follows: Decision Tree 90.34%, Naive Bayes 91.54%, and Neural Networks 94.91%. Neural networks can predict heart disease better than other techniques. The Chi square method has also been applied to determine the difference between the expected and the observed results, and the proposed system proved its accuracy at 86.54%.

A Clinical Data Analysis Based Diagnostic Systems for Heart Disease Prediction Using Ensemble Method

“…Elshafie et al [8] used an existing approach on the ADNI dataset that optimizes the computational cost over other existing prediction methods. Other existing work [9] has higher computational cost in terms of time complexity as the number of parameters increases the time complexity of the model. In this work, we have deployed a combination of strategies that work together to accomplish a single goal: feature extraction from datasets, 3D-to-2D grayscale picture conversion, and image size reduction by clipping [10] .…”