摘要
本文将格值信息论应用于医学专家系统中,重点解决医学专家系统中的两大问题:不精确推理和阈值技术。为此,首先引进格值信息代数,来论证经验公式的合理性;其次引进格值信息熵来正确应用阈值技术。
The two main characteritics in medicine-expert systems are inexcat deduction and thershold technique.In inexact deduction,many experience formulae are used and it is difficult that we use classical mathe- matics or fuzzy mathematics to prove them.So,Lattice-Valued Information Algebra is introduced,which provides the theoretic basis for proving these formulae.It is well-know that threshold technique is used in most medi- eine-expert Systems.But in this technique,the linear-addition of score will cause difficulty when we select threshold values and the augmentation of error-diagnosis.Because of these,Lattice-Valued Information Entropy is presented,which can solve these problems satisfactorily.
出处
《沈阳理工大学学报》
CAS
1989年第2期15-21,共7页
Journal of Shenyang Ligong University
基金
国家自然科学基金会资助
资助号为6875039.
关键词
人工智能
信息论
概率逻辑
artificial intelligence
information theory
probability logic