Enabling Digital Trust for Predictions Made by AI in Heart Disease Prediction

Heart disease is a major health problem and can lead to illness and death if not detected and treated early. The health-care systems are flooded with numerous data. Especially, with the advancement in the new technologies and techniques in analytics, it is now possible to analyze any clinical data and derive useful information from the huge data. From data collection to decision-making, there is a significant scope for improvement. The main objective of this project is to design and develop a model using predictive data mining technique with the help of explainable AI to identify patterns and correlation that would be difficult or impossible to detect using traditional analysis methods.This paper aims to take advantage of a hybrid approach that is combining a predictive data-mining technique with rule-based system to predict. There are some of the prevalent techniques that serve the purpose of analyzing the data such as Neural Networks, Naive Bayes classification, Random Forest Classifier, KNN etc. From all these techniques, the technique which will give the highest accuracy score will be considered for further implementation of model that is in our case it’s Artificial neural network with 98% accuracy. This paper benefits from ANN with rule-based system to reaffirm and then decide, whether the prediction results of heart disease can be trusted or not. The integration of ANN with rule-based system gave an accuracy of approximately 90%. Real-world data from health-care organizations must be gathered, and all available techniques must be compared to find an appropriate method with maximum accuracy. The ultimate goal is to use data mining techniques with explainable AI to transform the health-care industry into one that is truly knowledge-rich and highly effective.