Particle swarm optimization algorithm for parameter estimation in
Gamma-Poisson distribution model of k-tree distance
Feixia Lu
Dingyuan Mo
Meng Gao
School of Mathematics and Information Science, Yantai University, Yantai, 264005, China
Key Laboratory of Coastal Environmental Processes and Ecological Remediation, Yantai Institute of Coastal Zone Research,
Chinese Academy of Sciences, Yantai, 264003, China
Univerisity of Chinese Academy of Sciences, Beijing, 100049, China
Computational Ecology and Software
ISSN 2220-721X
http://www.iaees.org/publications/journals/ces/online-version.asp
2015
5
4
276
285
International Academy of Ecology and Environmental Sciences
Hong Kong
25 July 2015
3 August 2015
1 December 2015
spatial point pattern
distance sampling
point to tree distance
density estimator
Distance sampling is a flexible and efficient inventory technique in forestry and ecology, especially in highly dense plant communities, and in difficult terrain. Point-to-tree distance or tree-to-tree distance was used to estimate characteristics of the spatial point pattern mapped from particular spatial locations of plant or tree individuals. For random spatial point patterns, there is an ideal probability distribution model of point-to-tree distance resulting in unbiased density estimators. For aggregated spatial point patterns, Gamma-Poisson probability model of point-to-tree distance corresponding to Gamma-Poisson point process is one candidate model. Although the density estimator based on Gamma-Poisson model is biased, it performs satisfactorily in practical applications. However, numerical method to compute the maximum likelihood estimates of Gamma-Poisson model is very complicated. In this paper, a parameter optimization method, particle swarm optimization algorithm, is applied for parameter estimation in Gamma-Poisson model. The results showed that the new parameter estimation method was efficient and not constrained by the sample size; therefore, the computational complexity was significantly reduced. We suggest this parameter optimization for density estimation in forestry and ecology.
DOI 10.0000/issn-2220-721x-compuecol-2015-v5-0020
http://www.iaees.org/publications/journals/ces/articles/2015-5(4)/particle-swarm-optimization-algorithm-for-parameter-estimation.pdf