Estimation on the Number of Shared Species via a Jackknife Procedure

在此計劃書中︽我們將針對兩群落中的共同種類數提出一系列摺刀估計量〈依據初步的研究︽各階的估計方法的原理相當簡單︽而且估計量的形式可以完全寫出︽因此計算上相當容易〈理論上︽高階摺刀估計量雖可消除較多估計量的偏誤︽但是相對的高階摺刀估計量也會伴隨較高的統計變異︽因此應用上︽本計劃書也會提出一套假設檢定的方法來決定摺刀估計的階次〈由於文獻上已經有一些共同種類數的估計方法︽為了評估這些估計方法︽以及測試我們所新提出的方法︽本計劃書考慮用兩個實際普查的森林資料當成抽樣母體︽這兩筆資料接來自馬來西亞︽因此在氣候或地理位置上都有較多的共同性︽相信這樣的模擬設計應該會比較符合真實的情況〈In this proposal, we will present a sequence of jackknife estimators for the numberof shared species in two communities. With simple and explicit formulae, theseestimators are supposed to be computed easily. As the jackknife orderincreases, the resulting estimator is expected to gain a smaller bias yet isusually accompanied with a larger variance, thus a sequential testing procedurewill be proposed to determine the order in applications as well. Several typicalmethods will be evaluated their performance through a Monte Carlo simulationstudy. To mimic data sampled from the real world, two forests census datasetsestablished in Malaysia and with 209 shared species between them wereconsidered to be our sampling population