Neural network paradigm comparisons for appendicitis diagnoses
Russell C. Eberharta b, Roy W. Dobbinsa b, Larrie V. Huttona b
- aApplied Physics Laboratory, Johns Hopkins University, Laurel, MD, USA
- bAppl. Phys. Lab. Johns Hopkins Univ., Laurel, MD, USA
Papers open inside SciRah from the original publisher or repository; SciRah never hosts them.
Abstract
The results of comparisons among diagnoses of appendicitis versus nonspecific abdominal pain using three neural-network paradigms are reported. The paradigms used were the back propagation, binary adaptive resonance theory, and fuzzy resonance paradigms. It appears, from the limited testing done, that the back-propagation network performs best. Also discussed is the need to standardize input data files to facilitate paradigm comparisons and minimize software system development time. A structure for network input data files that could contribute to a process of standardization is proposed. The work is part of an effort to develop a medical practice support system to be used in isolated environments such as submarines.>