应用模糊数学建立甲状腺疾病鉴别诊断系统的研究  

Development of a Thyroid Disease Differential Diagnosis system Using Fuzzy Math

作　　者：舒柏华[1] 叶维新[1] 张永学[1] 安锐[1] 

机构地区：[1]同济医科大学环境医学研究所环境监测教研室,同济医科大学实验医学研究中心核医学研究室,同济医科大学附属协和医院核医学科

出　　处：《同济医科大学学报》1995年第4期311-313,315,共4页Acta Universitatis Medicinae Tongji

摘　　要：在统计大量有关甲状腺疾病的临床资料以及有关临床医生经验的基础上、应用模糊数学方法建立甲状腺疾病鉴别诊断系统。数学模型采用模糊关系方程。算子为归一化后有界求和与Ｚａｄｅｈ算子的结合。这样可起到互补作用，提高诊断水平。然后根据最大隶属度原则进行鉴别诊断。有些症状或指标能决定思维方向，此时采用判别树的方法进行鉴别诊断。当仅根据症状及体征作甲状腺功能分类时，符合率为９１．１％，在取得必要的理化检验结果后，进行临床诊断时，符合率为９７．２％。On the basis of a great amount of the clinical data of thyroid diseases and doctors'experience the author developed a thyroid disease diagnosis system by using fuzzy math.The model is a fuzzy relation equation.The operator is a limited sum of unification and Zader's operator,In this way,they have the function to compensate each other and can enhance diagnosis. The system achives differential diagnosis by means of the principle of maximum membership degree. Some symptoms or indexes can determine the direction of thinking,and in this case,we adopt distinguishing tree method in diagnosis.The system's agreement with expert diagnosis reached 91.1% in 257 cases and when it classifies thyroid function only on the basis of symptoms and indexes. If necessary laboratory results are available,the agreement rate was 97.2%.

关 键 词：甲状腺疾病 诊断 计算机 模糊数学 鉴别诊断系统 

分 类 号：R581.04[医药卫生—内分泌] R447[医药卫生—内科学]