Expert System Of Decision Tree Based On Rough Sets Theory And Its Application In Automatic Recognition Of ECG

With the development of artificial intelligence, expert system has been the vitallest leading edge in the intelligent information-processing field and computer science field. The expert systems with the function of machine learning or data mining, which can extend the approaches to obtaining knowledge, have attracted more attention in recent years. Decision tree learning is one of important branches in inductive learning and it can make the auto-obtaining of knowledge come true.This paper applies expert system of decision tree based on rough sets to the basic research of recognition and diagnosis of heart disease, and expatiates the basic design idea and the process of learning algorithm.Rough sets theory can deal with uncertainty knowledge, and educe the decision-making and classific rules through knowledge reduction. This paper brings forward a new method for knowledge reduction based on relation matrix. This method can turn the problem about reduction into set covering problem, and it has great potential to resolve the reduction problem in rough sets in the point of view.At present, there are some algorithms combining rough set and decision tree, which excel the series algorithms of ID3 in some aspects, but their combination methods are still not perfect. Against the deficiency of the decision trees in rough sets based on the differentiation degree, we improved the algorithm and the results of experiments proved its advantage.The improved algorithm of decision trees based on rough sets, which is applied to the expert system of heart disease diagnosis, improves the agility, practicability and alternation of the system, saves time for diagnosis and economizes memory spaces for rules.