Confidence bounds for the probability of unobserved species
Several methods for creating confidence bounds for estimators used in the species problem were examined. An interval based on the work of Robbins,$$V\sb1 + 1.64 \sqrt{.61\over n + 1},$$where $V\sb1$ is the number of singletons, was the best 95% bound for the probability of an unobserved species amon...
Saved in:
| Main Author: | |
|---|---|
| Format: | Dissertation |
| Language: | English |
| Published: |
ProQuest Dissertations & Theses
01-01-1991
|
| Subjects: | |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Several methods for creating confidence bounds for estimators used in the species problem were examined. An interval based on the work of Robbins,$$V\sb1 + 1.64 \sqrt{.61\over n + 1},$$where $V\sb1$ is the number of singletons, was the best 95% bound for the probability of an unobserved species among the ones studied here. The consequences of using the jackknife and the bootstrap to create confidence bounds were also examined. Analytic results were obtained for the expectation, bias and variance of the jackknife. Although it was shown that the jackknife produces biased estimates, Monte Carlo studies indicate that, under some circumstances, a jackknifed estimator will have a smaller mean square error than the unbiased estimator. Other simulation studies found that the bootstrap performs poorly in this problem. |
|---|---|
| ISBN: | 9798209057604 |