Autoregressive sequences and morphometric distributions
Mitolo, V.; Manodoro, V.; Saccia, M.; Maffione, A.B.; Bosco, L.
Bollettino della Societa Italiana di Biologia Sperimentale 64(5): 403-410
1988
ISSN/ISBN: 0037-8771 PMID: 3179022 Document Number: 326192
In morphometric estimates often one considers the occurrence of a given detail per unit surface or volume: e.g., the number of the cells in mitosis per unit surface in a microscopical preparation. When the state of a cell (e.g: mitotic or intermitotic) is not correlated with the state of the neighbouring cells, the statistical distribution of the observations is binomial. When the state probability of a cell is correlated (positively or negatively) with the state of the immediately neighbouring cells, the distribution departs from the binomial model: in particular, a positive correlation determines distributions with a comparatively higher variance while a negative correlation brings about distributions with a comparatively reduced variance.