Publication at Third Faculty of Medicine |
1994
Abstract
To assess the potential for applying artifical neural networks to the identification of Gram-negative rods, four different networks were trained, differing by the number quality of strains in the training and test sets. A network was trained using 3429 strains and tested with 6859 strains identified 95.5% strains at the species level and 97.7% strains at the genus level.
The sensitivity was 95.5 and specificity 99.9%. It was concluded that artificial neural networks are useful tools for phenotypic identification of Gram-negative rods