Abstract: Constructing and using intelligent computer systems to assist human experts to solve problems of disease diagnosis about people and livestock is one of hot spots in the field of medical artificial intelligence research.Traditional diagnosis systems taking no consideration of timing and associability between clinical characteristics are always with the high rate of misdiagnosis，and lack of practicality.At the aim to improve the efficiency of machinery-diagnosis，in case of goats，the clinical relevance of the disease is analyzed，and mechanisms of associated factor into the method of Certainty-Factor-knowledge-representation is indroduced；and the CF-value of associate-factors with big weights is predicted through the channel of forecast by lowest-value，highest-value，weighted-value.Experimental results show that this method can significantly the improve inference rule activation rate and reduce the rate of misdiagnosis.

Key words: bigger-weighted-related, Certainty Factor forecast, weighted uncertainty reason, disease diagnosis, Certainty Factor Integration

摘要： 设计利用智能计算机系统协助人类专家求解人畜疾病诊断问题是医学人工智能的研究热点之一。传统诊断系统没有处理临床之间的时序相关性，漏诊率高，实用性低。为改进机器诊疗效率，以山羊为例，分析了疾病临床的关联性，在可信度知识规则表示方法中引入关联因子机制；用低值联想，高值联想，加权联想等算法修下重正关联因子可信度。实验表明：重权关联因子可信度联想显著地改进了推理规则激活率，大大降低了漏诊率。

关键词: 重权关联因子, 可信度联想, 加权不确定推理, 疾病诊断, 可信度合成