A Hybrid Machine Learning and Association Rule Mining Approach for Accurate Heart Disease Prediction in Distributed Healthcare Databases

Abstract

Heart diseases continue to be one of the leading causes of death globally, which is why early diagnosis remains highly important for improved patient outcomes as well as overall less expensive healthcare delivery. Traditional diagnostic methods rely upon clinical investigations, which can include ECG, analysis of cholesterol, and various imaging techniques, but these might not take full advantage of large and complex medical datasets. This study develops a hybrid model for heart disease predictions that combines association rule mining with machine learning approaches on distributed medical databases. The model looks to analyze patient health records, lifestyle variables, and clinical attributes to uncover hidden associations between risk factors and heart disease occurrence. First, the data is manipulated and cleaned with respect to such an attribute as normalization, removing noisy instances in the dataset, and feature scaling before it is split 80/20 into training set and test set to ensure a reliable evaluation of potential models. Association rule mining using the Apriori algorithm yields meaningful relations involving clinical attributes such as chest pain type, exercise-induced angina, thalassemia, and gender. Rules derived from this analysis would further enhance the prediction capacity of the machine learning classifiers. An ensemble learning technique with Adaptive Boosting (AdaBoost) was implemented to improve the prediction capability and classification reliability of the proposed system. Experimental assessments extracted from the proposed system resulted in excellent predictive performance with 98.2%, 99.9% specificity, and an overall accuracy estimate of 98.1% performance, indicating strong reliability in diagnostics.The conclusion now drawn is that the integration of association rule mining with ML in distributed healthcare resources has indeed boosted the accuracy of heart disease predictions. It would support healthcare workforces in timely detection, decision support, assisted personalized treatment planning in a new wave-a chance to contribute toward advanced-augmented patient care and preventive health-care systems.