Predictivity in taxonomy and the probability of a tree

The contribution of J. S. L. Gilmour to numerical taxonomy is reviewed. His important concept of natural classification, as being general-purpose classifications with high predictivity, led to the development of ideas of information content, unit characters and equal character-weighting. The concept...

Full description

Saved in:
Bibliographic Details
Published in:Plant systematics and evolution 1989-01, Vol.167 (1/2), p.43-57
Main Author: Sneath, P. H. A.
Format: Article
Language:English
Subjects:
Computer networking
Datasets
Decision trees
Information content
Numerical taxonomy
Phylogenetics
Plants
Random sampling
Topology
Trees
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The contribution of J. S. L. Gilmour to numerical taxonomy is reviewed. His important concept of natural classification, as being general-purpose classifications with high predictivity, led to the development of ideas of information content, unit characters and equal character-weighting. The concept of predicitivity is extended to taxonomic trees (phenograms or cladograms). Under certain assumption of random sampling of characters it is shown that the probability of recovering the correct tree topology or tree-form may be small if characters are few. There may be very many topologies or tree-forms, every one of which has individually a low probability. It is, however, possible to estimate the aggregate probability of trees which have more than some specified resemblance to the "correct" tree. The practical prospects of estimating the distribution of tree probabilities are discussed.
ISSN:0378-2697
1615-6110
DOI:10.1007/BF00936546