Abstract: Diagnosis efficiency of disease diagnosis expert system is affected by its bottleneck of knowledge-rule-explosion and invalid combinations which are built on the base of production rule. In order to help doctors diagnose disease, the theoretical model of disease-resemblance pattern recognition based on certainty factors vectors and fuzzy membership factors vectors, its data structure model, its computer model recognition algorithm and its practice methods are proposed. Experimental data shows that compared with the individual human expert, the model can obtain higher diagnostic accuracy rate of 80%, effectively reduce the misdiagnosis rate, and carry a preferential-comprehensive diagnosis performance.

Key words: disease pattern, recognition, disease diagnosis, resemblance

摘要： 产生式规则专家系统因规则组合爆炸与无效匹配等问题影响诊断效率。为协助人类专家求解疾病诊断问题，分析模式识别与该问题解决手段的相关性和相似性，提出基于可信度向量和模糊隶属度向量的人工智能疾病模式识别理论模型、数据结构模型、计算机模式识别算法与实践方法。现场实验数据显示，与人类专家个体相比，该模式能获得80%的诊断正确率，有效地降低误诊率，具备较好的综合诊断性能。

关键词: 疾病模式, 识别, 疾病诊断, 相似度

CLC Number: 

• TP181