Monitoring the richness and the diversity of species living in an ecosystem is an important goal of ecology. To this purpose, measures of biodiversity have been introduced as statistical summaries of the abundance vector. In particular, we take into consideration the Gini–Simpson and the Shannon–Wiener indices, along with the effective number of species calculated through these measures, proposed, respectively, by Laakso and Taagepera (Comp Polit Stud 12:3–25, 1979) and Leti (Statistica descrittiva, Bologna, Il Mulino, 1983). It is an open question how to associate to these indices a measure of uncertainty. In this paper we compare confidence intervals based on these measures, calculated through three different bootstrap methods: percentile, -t and accelerated bias-corrected percentile. We recommend to practitioners to use the percentile procedure, as it is straightforward and computationally feasible, providing results very close to those obtained by more complex techniques.
Pesenti, N., Quatto, P., Ripamonti, E. (2017). Bootstrap confidence intervals for biodiversity measures based on Gini index and entropy. QUALITY & QUANTITY, 51(2), 847-858 [10.1007/s11135-016-0443-x].
Bootstrap confidence intervals for biodiversity measures based on Gini index and entropy
Pesenti, N;QUATTO, PIEROSecondo
;RIPAMONTI, ENRICOUltimo
2017
Abstract
Monitoring the richness and the diversity of species living in an ecosystem is an important goal of ecology. To this purpose, measures of biodiversity have been introduced as statistical summaries of the abundance vector. In particular, we take into consideration the Gini–Simpson and the Shannon–Wiener indices, along with the effective number of species calculated through these measures, proposed, respectively, by Laakso and Taagepera (Comp Polit Stud 12:3–25, 1979) and Leti (Statistica descrittiva, Bologna, Il Mulino, 1983). It is an open question how to associate to these indices a measure of uncertainty. In this paper we compare confidence intervals based on these measures, calculated through three different bootstrap methods: percentile, -t and accelerated bias-corrected percentile. We recommend to practitioners to use the percentile procedure, as it is straightforward and computationally feasible, providing results very close to those obtained by more complex techniques.File | Dimensione | Formato | |
---|---|---|---|
BootstrapConfidenceIntervalsFo.pdf
Solo gestori archivio
Tipologia di allegato:
Publisher’s Version (Version of Record, VoR)
Dimensione
565.37 kB
Formato
Adobe PDF
|
565.37 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.