Experimental Methods in Computer Taxonomy

Open Access
Publisher Website
Google Scholar
Abstract
SUMMARY: In the preparation of taxonomic data for computer analysis, some new methods are proposed which compensate for certain inadequacies in present techniques without violating the essential Adansonian axiom that all properties of organisms are of equal significance in the creation of taxa. A modified scoring technique is presented for use with tests wherein two or more alternative responses (none necessarily negative) are possible. The proposed method might well be extended to the scoring of all data; one is never sure of the significance of a ‘negative’ result, and it could be considered that at least two alternatives of equal value in classification may exist for any determinable property of an organism. It is suggested that computed relationships among organisms might better be expressed in terms of distance, which is a logarithmic function of the similarity ratio in current use. A new diagnostic parameter, P, is defined as a quantitative estimate of the proportion of organisms in any group (taxon) which possess a given property. Computer programs based on such values could be devised for diagnosis of organisms whose identitv is unknown.

This publication has 9 references indexed in Scilit: