New system forecast e-health by using hybrid of association rules and decision tree

Abstract

E-health technologies can play a major role in improving the lives of patients and especially with chronic diseases, these technologies can provide quality care at a distance without consulting their doctor regularly. These technologies provide many benefits for all stakeholders: patients and health providers to patients, the main advantage of e-health is remote access of this information, and communications services as well as communication. Directly with professionals and health services without having to travel to doctors. For providers of health, the main advantages of e-health is improving access to real patient data remotely and improve the quality of their decision and their services through systems help in the decision. We propose in this paper to compare the two techniques of data mining association rules and decision trees in discussing their advantages and limitations depending on the efficiency, accuracy and timeliness in real time. Our discussion leads us to propose a new solution of a system of monitoring and suggestion based on the hybrid use of association rules and decision trees where the association rules in first phase is used to generate the relevant relationships for a homogeneous group of patients who share the same profile and the second phase involves generating a decision tree for each group of patients, from medical relationships that are generated in the first phase and the elimination of transitivity and regenerating the canonical decision tree