Risk Prediction for Coronary Heart Disease Using C5.0 Decision Tree Algorithm

Abstract

Heart Disease is now-a-days one of the most leading causes of death rate in the worlds. Here we aim to explain the CHD predictive model using C5.0 one of the DT algorithm. In a every single day there are huge amount of data were providing in the hospitals. For making a correct decision for the disease some of the hidden information are not mined properly or effectively. So, here we are using some DM techniques which can make a solution to this situation. This work has been developed using the DM techniques namely DT (CART, C5.0) algorithms and classification (Naive Bayes, Neural Network) algorithms. Result shows each technique has its own specific strength in realizing the objective of the defined mining goals. Using attributes (age,sex,cp, trestbps, chol, fbs, restecg, thalach, exang, oldpeak, slope, thal, ca, num) it can predict the likelihood of patients getting heart disease. In this paper we studied and validated the predictive power of DM algorithms by comparing the performance of C5.0 with two classifier algorithm Naive Bayes, Neural Network and one DT algorithms CART. At finally we are comparing it Accuracy, Sensitivity, Specificity with two different types of classification algorithm and one DT algorithm. Contrary to the former study, the C5.0 algorithm performed best than the other two classifiers algorithm and one DT algorithm in predicting CHD, and C5.0 have the highest predictive power. This paper provides an insight about C5.0 DT algorithm used to predict the heart diseases.