摘要： 传统疾病诊断专家系统通过单次推理过程进行知识运用，其知识利用率低，结论准确度低且不具备对比度。该文以山羊为例，运用面向对象的知识表示方法对疾病诊断领域知识库进行建模，结合专家训练过程与诊断模型，提出知识库多引擎判决思想，多次牵引知识库运用知识，构造面向对象加权不确定判决和样本匹配判决算法。实验结果表明，多引擎诊断模型提高了知识库数据资源的利用率，改善了诊断准确度，增加了对比度。

关键词: 样本匹配, 加权不确定判决, 重权关联因子, 知识表示

Abstract: The traditional disease diagnostic expert system only has a single reasoning process to use knowledge, so that its knowledge utilization is inefficient. Its conclusion has low accuracy and without contrast. This paper makes example of goat, modelings disease diagnostic knowledge base by the use of object-oriented knowledge representation. According to the training process and diagnostic model, it proposes the ideology of multi-engine ruling based on one knowledge base, multi-tract to operate knowledge, constructing the object-oriented weighted uncertain judgments and decisions basis on samples matching algorithm. Experimental results show that multi-engine diagnostic model can improve the utilization rate of data resources in knowledge base, improve the accuracy of diagnosis, increase the contrast.

Key words: samples matching, weighted uncertain judgments, bigger weight associated factor, knowledge representation

中图分类号: 

• TP311