|
Activity Number:
|
377
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #305925 |
|
Title:
|
Estimating the Species Richness by a Poisson-Compound Gamma Model
|
|
Author(s):
|
Ji-Ping Wang*+
|
|
Companies:
|
Northwestern University
|
|
Address:
|
2006 Sheridan Road, Evanston, 60208,
|
|
Keywords:
|
species richness ; Poisson mixture
|
|
Abstract:
|
Suppose $D$ distinct species are observed from a infinite population consisting of total $N$ (unknown) species. It has been well recognized that $N$-estimate is unbounded if infinitely rare species are possible. Here we consider to estimate the number of species whose abundance is above a threshold. We model the probability of observing $X$ individuals from each species by a Poisson-mixed Gamma model, i.e. $f[X;Q(\lambda)]=\int\frac{e^{-\lambda}\lambda^x}{x!}dQ(\lambda)$ where $Q$ itself is a mixed Gamma. The Gamma is estimated by nonparametric maximum likelihood method. Instead of estimating $N$, we estimate $\sum I(\lambda_i>\lambda_0)$ based on the posterior distribution of $\lambda|X$. Confidence interval is constructed based on a resampling procedure. This method is illustrated using simulated and real data.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2006 program |
