Refine
Year of publication
- 1983 (1)
Document Type
- Article (1)
Language
- English (1)
Has Fulltext
- yes (1)
Is part of the Bibliography
- no (1)
Two sets of simulated, randomly distributed, vegetation data of different complexity (3 and 50 species) were analysed with both a "single linkage" and a "sum-of-squares" clustering method. The 3-species data showed a strong clustering with both methods. Only the "sum-of-squares" method showed strong clustering in the data with 50 species. The "single linkage" clustering resulted in "chaining", pointing to a more continuously variable vegetation. It is emphasized that many clustering methods result in strong clustering even when the data are random.