Current issue: 58(5)
Select issue
Silva Fennica 1926-1997
1990-1997
1980-1989
1970-1979
1960-1969
1950-1959
1940-1949
1926-1939
Acta Forestalia Fennica
1992-1999
1984-1991
1974-1983
1968-1973
1953-1968
1933-1952
1913-1932
Articles containing the keyword 'maximum likelihood method'
Category : Research article
article id 956,
category
Research article
(2013).
Evaluation of the double normal distribution for tree diameter distribution modeling.
Silva Fennica
vol.
47
no.
2
article id 956.
https://doi.org/10.14214/sf.956
Keywords:
diameter distribution;
simulated annealing;
goodness-of-fit;
Weibull distribution;
maximum likelihood method;
method of moments
Abstract |
Full text in HTML
|
Full text in PDF |
Author Info
The double normal distribution consists of two normal distributions truncated at their means and then combined in such a way, that points of truncation now become the overall distribution mode. So far, parameters of the double normal distribution have been estimated exclusively using the method of moments. This study evaluates the maximum likelihood method for the estimation of the double normal distribution parameters in Scots pine stands in Poland, and compares it to the results of the method of moments and the two-parameter Weibull distribution fitted using the maximum likelihood method and the method of moments. Presented results show that it is not recommended to use the maximum likelihood method of parameter fitting with Nelder-Mead and quasi-Newton optimization algorithms for the double normal distribution for small samples. However, it can be used for large samples, giving the fit comparable to the two-parameter Weibull distribution and providing parameters having sound practical and biological meaning. In the case of smaller samples for the double normal distribution it is recommended to apply the maximum likelihood method with the alternative simulated annealing optimization algorithm, use the method of moments or substitute the described distribution with more the flexible and robust Weibull distribution. For the smaller samples, the method of moments was superior to the maximum likelihood method.
-
Zasada,
Warsaw University of Life Sciences, Faculty of Forestry, Laboratory of Dendrometry and Forest Productivity, Nowoursynowska 159, 02-776 Warsaw, Poland
E-mail:
michal.zasada@wl.sggw.pl
Click this link to register to Silva Fennica.
If you are a registered user, log in to save your selected articles for later access.
Your selected articles
Silva Fennica ·
Finnish Society of Forest Science
