基于逻辑的乳腺癌诊断与治疗专家系统

本系统收集了天津市领域专家的经验,由编译性人工智能语言TurboProlog实现,共由七个模块-知识率、推理机构、解释机构、半自动人工知识获取、知识库管理、动态数据库和人机接口组成。本系统利用产生式规则表达知识,知识库中有多达1242条规则。在推理机构中,采用两类策略:启发式正向和反向推理方法,并利用黑板结构作为公共动态数据区。。提供了半自动人工知识获取以能获取新知识、丰富和更新知识库。在黑板上保留的推理进程是解释机构的素材。总系统共占用118KB内存,在IBM PC AT微机上运行正常,起到了总结和推广领域专家诊断与治疗乳腺癌经验的作用。

Logic Based Diagnosis and Treatment ExpertSystem for Breast Cancer

The system collects the domain experts' experiences of Tianjin to be realized by compiled artificial intelligent language Turbo Prolog, Which is constructed of seven model blocks called knowledge base, reasoning mechanism, explanation mechanism, semiautomatic manual knowledge acquisitiion, management on knowledge base, dynamic data base and man machine interface. This system uses productive rules to represent knowledge. In knowledge base there are more than 1242 rules. In reasoning mechanism, two kinds of strategy: heuristic forward and backward reasoning method are adopted, and blackboard structure which is a common dynamic data area being applied. The semiautomatic manual knowledge acquisition is provided, in order to get the new knowledge to enrich or renew the knowledge base. The reserved reasoning process on the blackboard is the material for explanation mechanism. The total system occupies 118kb memories and works well at IBM PC AT, which is taking a part of summing and extending for domain experts' experiences in the diagnosis and treatmnet for breast cancer.