透過您的圖書館登入
IP:13.59.122.162
  • 學位論文

直線穿越抽樣法有乘法性測量誤差時之近似估計方法

The approximate estimations in the line transect model when the covariate is subject to multiplicative measurement error

指導教授 : 黃逸輝

摘要


在生態學中,調查對象的密度或總數是相當重要的參數,而距離抽樣則是方便且常採用 的抽樣方法,本文討論的距離抽樣是直線穿越抽樣法,並且是觀測距離無法精準量測而有測量誤差時的統計分析問題。在直線穿越抽樣法有乘法性測量誤差時,本文將乘法性測量誤差模型作自然對數轉換,使轉換後的模型為相加性測量誤差模型,並比較相加性測量誤差時常見的修正方法:迴歸校正(regression calibration)、泰勒(Taylor)展開式及模擬外推法(simulation extrapolation)等的電腦模擬表現。

並列摘要


In ecology, the density or the total number of residents are very important parameters. To estimate these parameters, the distance sampling is a convenient and often used sampling methods. We discussed this sampling method in line transect experiment when covariates are subject to multiplicative measurement errors. We evaluate the performances of methods of regression calibration, Taylor approximation and simulation extrapolation by simulation studies.

參考文獻


Fan, K. S. (2009). Measurement Errors and Imperfect Detection Rates on theTransect Line in Independent Observer Line Transect Surveys. Journal of Data Science 7, 331-347.
Marques, T. A. (2004). Predicting and Correcting Bias Caused
by Measurement Error in Line Transect Sampling Using Multiplicative Error Models. Biometrics, 60, 757-763.
Chen, S. X. (1998). Measurement Errors in Line Transect Surveys. Biometrics 54, 899-908.
Chen, S. X. (2001). Measurement Errors in Line Transect SurveysWhere Detectability Varies with Distance and Size. Biometrica 57, 732-742.18

延伸閱讀