Dados Bibliográficos

Resumo

The statistical method of nearest neighbor analysis is presented for the study of distributional patterns of artifacts over occupation floors. It is compared with the previously presented method of dimensional analysis of variance. Nearest neighbor analysis is found to be much more sensitive in its detection of non-random spatial clustering. It has the advantage of not being particularly limited in application by problems of size or shape of the area under study, although it does require coordinates for each artifact and cannot be applied when only counts per grid unit are known. On the other hand, nearest neighbor analysis encounters considerable problems in defining the artifact clusters on an area and in comparing the distributions of several artifact types. These problems severely limit the utility of nearest neighbor analysis at the moment. Dimensional analysis of variance handles them better.

Referências Citadas