Peyret M, Flandrois J P, Carret G, Pichat C
Université Claude-Bernard, CNRS Unité Associée 243, Centre Hospitalier Lyon-Sud.
Pathol Biol (Paris). 1989 Jun;37(5 Pt 2):624-8.
An expert system (cadi-yac), written in Turbo-Prolog and working on IBM PC and Bacanal + (a management software of microbiology laboratory) was used to recognize and correct the phenotype of antibiotic sensibility. The results were there of API ATB system. The knowledge was adapted from two references works. A routine use of the expert system give a correct recognition of enzymatic profile in more than 80% cases for the beta-lactams and more than 98% cases for the aminosides. The mistakes detected by cadi-yac, were often interpreted as deficiency of API system by humans experts. The expert system mistakes (1.5%) were due composites phenotypes.
一个用Turbo - Prolog编写、运行在IBM PC和Bacanal +（微生物实验室管理软件）上的专家系统（cadi - yac）被用于识别和校正抗生素敏感性表型。结果来自API ATB系统。知识取自两部参考文献。专家系统的常规使用能在超过80%的β-内酰胺类病例和超过98%的氨基糖苷类病例中正确识别酶谱。cadi - yac检测到的错误，人类专家常常将其解释为API系统的缺陷。专家系统的错误（1.5%）是由于复合表型。
