Identification of unknowns within a probabilistic system: The diagnostic value of attributes

Abstract

Using a data base underpinned by probability considerations in which a variety of attributes, some of which may be quantitative, are recorded for a number of “operational taxonomic units” (OTUs), a key system is described by which an unnamed specimen may quickly be identified. The concept of “diagnostic power” is introduced, by which each attribute is evaluated in terms of its potential contribution to identifying the unnamed specimen. Besides coverage of different types of attributes and the introduction of “diagnostic power”, the system has the advantages of incorporating multiple values of an attribute for each OTU and offering short-cuts to identification.

Keywords::

• Identification
• diagnosis
• probability
• classification
• taxonomy
• Drosera

Acknowledgements

I would like to thank my colleagues Dr A. Koenders and Dr K. Lemson for valuable criticism of this paper in draft and Dr P.L. Nimis and Dr S. Martellos for reading and commenting on the last version of the paper. Finally, I wish to express my appreciation to my friend Enrico Feoli, of Trieste, who undertook to see this paper through the publication process for me.

Notes

^1. OTU: “Operational Taxonomic Unit” – the items classified under this system.