The Application And Research Of Rough Sets And Evidence Theory In Medical Intelligent Diagnosis System

With the popularity of medical information and the application of a variety of medical testing equipment, leading to unprecedented growth of medical data which is characterized by incomplete, uncertain and redundancy. It is inefficient only by simple human refining and the use of the information. Therefore comprehensive development and utilization of a large amounts of data information using advanced computer and information processing technology, to achieve high diagnostic accuracy of intelligent medical diagnosis system, which has become one of the important research direction of the development of medical industry.Rough set theory is a data analysis method that can deal with imprecise, uncertain and incomplete information. The capability of good data processing is its greatest feature. But the rough set theory can not accuratly decision-making under various rules. And the evidence theory is an uncertainty reasoning theory, its handling of evidence comes from experts. It has a significant advantage to solve the problem of uncertainty in data fusion. According to this, the paper proposes a synthesis method based on rough sets and decision rules of evidence theory, and researchs the application in medical diagnostic system.The main researches and contributions are as follows:1. Analysis of the characteristics of clinic data, classification of a large number of diseases diagnosis information by the introduction of rough set theory. A dynamic attribute reduction algorithm based on rough entropy is proposed, aim at the need of dynamically establishing disease diagnosis knowledge base and the existing problems of reduction of rough set. The algorithm uses the new object to amend the original reduction relation, recursively computes the information rough entropy after system change using the results of the original information rough entropy, to save calculation quantities and to improve the reduction efficiency.2. A synthesis method based on rough sets and decision rules of evidence theory is proposed, because of the limitations of rough set in rules decision-making. Combining with clinical medicine, a diagnostic model based on rough set and evidence theory is constructed. The decision rules in decision table is as identification framework of evidence theory, Calculating the reliability and importance of the rules, synthesising evidence using evidence theory for decision-making results of different rules, finally obtaining the results of diagnosis.3. Based on the above algorithm, take rheumatism for instance, a rheumatism specialist diagnosis system is designed. The general framework and design idea of the system are detailed, including the design of case database, diagnostic knowledge base and so on. System test results achieve the effect of assisting doctors to diagnose, and verify the feasibility and effectiveness of the proposed method.