A High Accuracy Integrated Bagging-Fuzzy-GBDT Prediction Algorithm for Heart Disease Diagnosis

Associated with high morbidity and mortality, heart disease has become a severe threat to peoples health throughout the world. The recent development of Internet of Things (IoT) and machine learning in e-healthcare have contributed to the monitoring, prediction and diagnosis of heart disease. Particularly, the heart disease prediction can effectively facilitate disease prevention, diagnosis and timely treatment. However, traditional prediction models are weak in accuracy and generalization. In this paper, we propose a high accuracy integrated prediction algorithm for heart disease diagnosis. The fuzzy logic and Bootstrap Aggregating (Bagging) algorithm based on Gradient Boosting Decision Tree (GBDT) algorithm are combined to process heart disease data and generate multiple weak classifiers. At first, we integrate the fuzzy logic with GBDT to reduce the complexity of data. Moreover, we develop the Fuzzy-GBDT model integrated Bagging algorithm to avoid the interference of sensitive points and achieve partial parallelism. The simulation results show the proposed Fuzzy-Bagging-GBDT algorithm improves the accuracy and recall of heart disease prediction compared with GBDT.