Probabilistic information processing: Implementation and evaluation of a Semi-PIP diagnostic system

✍ Scribed by David H. Gustafson; John J. Kestly; Robert L. Ludke; Frank Larson

• Publisher: Elsevier Science
• Year: 1973
• Tongue: English
• Weight: 1018 KB
• Volume: 6
• Category: Article
• ISSN: 0010-4809
• DOI: 10.1016/0010-4809(73)90071-2

✦ Synopsis

A computer-aided diagnostic system called Semi-PIP which uses subjectively estimated patient attribute-disease relationships and is sensitive to conditional dependencies is described and tested. This model uses the best of four tested methods to subjectively classify data into conditionally independent complexes. Its performance is compared to two models which ignore conditional dependencies, and to three physicians diagnosing on the basis of symptoms, physical signs and laboratory information. Analysis of the results indicates that Semi-PIP is superior to the other computer models, and almost as accurate as the majority physician diagnosis. P(Di/S,. a es,) = P(S, . . . &IDi)P(Di) f: P(SI * *. sm/Di) P(Di)' (1) Copyright