×

Geometric methods in mathematical statistics. (Slovak) Zbl 0706.62058

This is an expository paper the aim of which is to explain geometric interpretations of some statistical quantities and to discuss possible use of geometric methods in statistics, e.g. geometric interpretation of the least squares estimators in both linear and nonlinear regression models and of I-divergence. Some geometry of curved exponential families is mentioned.
The main references are D. M. Bates and D. G. Watts, J. R. Stat. Soc., Ser. B42, 1-25 (1980; Zbl 0455.62028), O. E. Barndorff- Nielsen, Res. Rep. No.106, Aarhus Univ. (1984), and S. Amari, Differential-geometrical methods in statistics (1985; Zbl 0559.62001).
Reviewer: M.Huškova

MSC:

62J02 General nonlinear regression
PDFBibTeX XMLCite
Full Text: EuDML