2007 Volume 6 Pages 43-56
In growth analysis, irregular statistical models are needed to detect some property in a forest stand, such as thinning effect, self-thinning effect or difference of growth patterns. The irregular models do not permit to use usual statistical theory, and then, the properties of them were discussed by each problem. Recently, however, unified statistical theory for the irregular models has been developed. In this paper, some tests based on the locally conic model, which is representative irregular model, are elaborated. In addition, differences from usual tests are explained, and the need of the irregular models is investigated in growth analysis.