Comparative Study of Human Expertise and an Expert System: Application to the Diagnosis of Child's Meningitis
✍ Scribed by Patrice François; Claudine Robert; Jacques Astruc; Pierre Begue; Jean-Claude Borderon; Daniel Floret; Bernard Lagardere; Eric Mallet; Jean-Claude Pautard; Jacques Demongeot
- Publisher
- Elsevier Science
- Year
- 1993
- Tongue
- English
- Weight
- 327 KB
- Volume
- 26
- Category
- Article
- ISSN
- 0010-4809
No coin nor oath required. For personal study only.
✦ Synopsis
We present a general method of statistical evaluation for expert systems, applied to a system for the diagnosis of child's meningitis. Fifty difficult clinical cases of child's meningitis were submitted to the system, to seven senior specialists and to seven young physicians. Multidimensional analysis of the diagnosis of the infection category reveals that the two groups of physicians separate naturally and that the system is located among the group of experts. The study of the agreement of the microbiological diagnosis and therapeutic advice shows that the advice of the two groups of physicians is significantly different and that the advice of the system is significantly closer to that of the experts. This result is confirmed by the study of therapeutic errors. This type of study allows one to classify the performance of the system among physicians having different levels of expertise without referring to an objective solution.
📜 SIMILAR VOLUMES
To provide readers of Applied Ergonomics with a selection of current ergonomics literature likely to be of direct practical value, abstracts are published selected from the collection held at the Ergonomics Information Analysis Centre. These abstracts are classified in a similar manner to the main a
Obtaining unknown NMR parameters from experimental spectra of exchanging systems is, from a mathematical point of view, associated with the solution of an inverse problem. Inverse problems are usually solved by iterative procedures and calculations may not always converge and the solutions are rarel