Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/96607
標題: 使用出現與否資料決定最佳抽樣區塊大小
Determining the optimal quadrat size using occurrence data
作者: 汪于棻
Yu-Fen Wang
關鍵字: 出現與否資料;卜瓦松分配;負二項分配;occurrence data;Poisson distribution;Negative Binomial distribution
引用: Bonham, C. D. (2013). Measurements for terrestrial vegetation. John Wiley & Sons. Hwang, W. H., & He, F. (2011). Estimating abundance from presence/absence maps. Methods in Ecology and Evolution, 2(5), 550-559. Hwang, W. H., & Huggins, R. (2016). Estimating Abundance from Presence– Absence Maps via a Paired Negative‐Binomial Model. Scandinavian Journal of Statistics, 43(2), 573-586. Kendall, M., & Stuart, A. (1977). The advanced theory of statistics. Vol. 1: Distribution theory. London: Griffin, 1977, 4th ed. Pettigrew, H. M., Gart, J. J., & Thomas, D. G. (1986). The bias and higher cumulants of the logarithm of a binomial variate. Biometrika, 73(2), 425-435. Ståhl, G., Ekström, M., Dahlgren, J., Esseen, P. A., Grafström, A., & Jonsson, B. G. (2017). Informative plot sizes in presence‐absence sampling of forest floor vegetation. Methods in Ecology and Evolution, 8(10), 1284-1291. Walter, S. D. (1975). The distribution of Levin's measure of attributable risk. Biometrika, 62(2), 371-372. Walter, S.D. (1976) The estimation and interpretation of attributable risk in health research.Biometrics, 32, 829–849. 孫義方. (2006). 森林生態學研究的新潮流–森林動態樣區. 23
摘要: 
出現與否資料是一個方便、便宜且常常被使用的方法,所以如何選擇最佳的區塊大小是一個很重要的問題。在本篇文章中,使用卜瓦松模型及其兩種修正法和負二項模型求最佳的區塊大小。本文用電腦模擬研究比較兩種模型的配適結果,並且用洛科羅拉多島(Barro Colorado Island)樣區內普查的資料,探討本文提供的方法的實用性與適切性。結果不論在電腦模擬研究或是實例分析,對於來自於卜瓦松或是負二項分配的資料,使用負二項模型配適效果良好,故最後建議不論資料來自於卜瓦松或是負二項分配皆可以使用負二項模型配適。

Occurrence data are widely used data formats when practically collecting quadrats since ecologists can only pay attention to whether species are present or absent in the sampled quadrats. For doing so, how to determine a suitable quadrat-size before sampling becomes a critical issue. Some previous works suggested estimating procedures based on a Poisson model; however, such a model is applicable to model species individuals displaying random patterns. For aggregate patterns that are common found in ecological studies, the present work suggests using a negative binomial distribution. Additionally, the Poisson model can be treated as a special case of the negative binomial distribution, consequently which can be extensively applied to various distributional-patterns. This dissertation carried out not only a simulation study but also a empirical test using the Barro Colorado Island (BCI) data to explore the practicability and appropriateness of all methods provided in the dissertation. Conclusively, regardless of the simulation study or empirical test, using the negative binomial model outperforms using the Poisson distribution.
URI: http://hdl.handle.net/11455/96607
Rights: 同意授權瀏覽/列印電子全文服務,2021-07-26起公開。
Appears in Collections:統計學研究所

Files in This Item:
File SizeFormat Existing users please Login
nchu-107-7105018024-1.pdf451.55 kBAdobe PDFThis file is only available in the university internal network    Request a copy
Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.