Neural Networks and their Possible Use in Computer-Assisted Diagnosis
Abstract
Hospital corpsmen stationed on submarines frequently must make medical decisions without consulting a physician. Since evacuation decisions are extremely crucial and may involve aborting a mission, Computer-Assisted Medical Diagnosis (CAMD) is provided to confirm the need for medical evacuation. The Abdominal and Chest Pain modules deployed use a Bayesian approach to estimate the probability of a medical diagnosis. A variety of other algorithms, including various statistical tools and expert systems have been used elsewhere for this purpose. Currently a new tool, the neural network model, is being considered for CAMD. The purpose of this paper is to describe the mathematics involved in simulating a simple neural network and to discuss its potential in connection with the Navy's CAMD program.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1989
- Accession Number
- ADA223917
Entities
People
- Antonio Gino
- Jean Dunbar
Organizations
- Naval Health Research Center