Proficient Prediction of Heart Disease Using Enhanced Backtrack Pruned Classification Algorithm and Effective Data Cleaning Technique

Abstract

ABSTRACT newlineData mining plays a major role in predicting the future. The history of data stored can be analyzed to predict the future trends and supports to make strategic decisions to improve the performance and profit of the businesses. Many researchers considered healthcare industry data to diagnose and predict the diseases to help patients and doctors in many ways. This research focuses on the analysis and prediction of heart disease. Heart disease is one of the most hazardous diseases to human beings which shows the way to death all over the world since 15 years. newlineMany types of researchers have been done with the techniques of knowledge discovery in various fields for heart disease prediction and have shown the acceptable levels of accuracy. There were no tools in real time for the purpose of analysis and prediction of heart disease in earlier stages. The decision trees are implemented for the analysis of different training and test dataset for prediction of heart disease. The classification algorithms like Naïve Bayes, ID3, C4.5 and the proposed Enhanced Backtrack Pruned C4.5 (EBPC4.5) algorithms are analyzed. It is experimented and analyzed by considering the two major drawbacks called attributes split criterion and the problem of overfitting. It is found that the performance of proposed EBPC4.5 classification algorithm provides the best solution among other algorithms. newlineThe datasets from the University of California, Irvine Machine learning database are collected, preprocessed and used for analysis and prediction of the disease with the above mentioned classification algorithms. Also a local dataset from Arcot Digital and Computerized Medical laboratory is used for implementation and prediction of heart disease. As the UCI Machine learning dataset and Local dataset were subjected to inconsistencies, they are preprocessed to remove inconsistencies and to improve the mining and prediction results. An Enhanced data preprocessing technique is used for preprocessing. It includes the Multi Layer Perceptron Neural ne