Abstract
A computer program was developed to identify anaerobic bacteria by using simultaneous pattern recognition via a Bayesian probabilistic model. The system is intended for use as a rapid, precise, and reproducible aid in the identification of unknown isolates. The program operates on a data base of 28 genera comprising 238 species of anaerobic bacteria that can be separated by the program. Input to the program consists of biochemical and gas chromatographic test results in binary format. The system is flexible and yields outputs of: (i) most probable species, (ii) significant test results conflicting with established data, and (iii) differential tests of significance for missing test results.
Full text
PDF
Selected References
These references are in PubMed. This may not be the complete list of references from this article.
- Dybowski W., Franklin D. A. Conditional probability and the identification of bacteria: a pilot study. J Gen Microbiol. 1968 Dec;54(2):215–229. doi: 10.1099/00221287-54-2-215. [DOI] [PubMed] [Google Scholar]
- Friedman R. B., Bruce D., MacLowry J., Brenner V. Computer-assisted identification of bacteria. Am J Clin Pathol. 1973 Sep;60(3):395–403. doi: 10.1093/ajcp/60.3.395. [DOI] [PubMed] [Google Scholar]
- Werner H., Rintelen G., Kunstek-Santos H. Eine neue buttersäurebildende Bacteroides-Art: B. splanchnicus n. sp. Zentralbl Bakteriol Orig A. 1975;231(1-3):133–144. [PubMed] [Google Scholar]