Measuring and comparing the accuracy of species distribution models with presence–absence data

By Canran Liu, Matt D. White, G. Newell. 2011

440 citations recorded by the source corpus.

Subfields: Ecological Modeling, Ecology, Nature and Landscape Conservation

View Paper: https://www.semanticscholar.org/paper/5779a18b7af949559841317d990fe83466e73f50

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Ecological Modeling

In Ecology

In Nature and Landscape Conservation

Citations in the corpus

Cross-validation of species distribution models: removing spatial sorting bias and calibration with a null model. (2012, methodology): “Key words: AUC; Bioclim; cross-validation; MaxEnt; model evaluation; niche model; pairwise distance sampling; spatial autocorrelation; spatial sorting bias; species distribution model.”
A comparison between Ensemble and MaxEnt species distribution modelling approaches for conservation: A case study with Egyptian medicinal plants (2020, background): “SDM has traditionally been focused on single-species presence data (Elith et al., 2006; Liu et al., 2011; Thuiller, 2003), but recent emphasis on the importance of community and biotic interactions in SDMs has propelled multi-species SDMs such as stacked SDMs (S-SDMs) or joint SDMs (J-SDMs) to the…”
On the selection of thresholds for predicting species occurrence with presence‐only data (2015, methodology): “For presence/absence data, the identities of both presence and absence data are known, and the related confusion matrix and calculation of various accuracy metrics can be found in many papers (Liu et al. 2011).”
Climate change threatens European conservation areas (2011, methodology): “The consensus was based on a weighted mean probability of occurrence per species and per grid cell, where weights are obtained from the TSS obtained on the evaluation data (Marmion et al. 2009).”
Mammal defaunation as surrogate of trophic cascades in a biodiversity hotspot (2013, background): “Sensitivity is the probability that the model correctly predicts an observation of a species at a given site, and specificity is the probability that the model correctly predicts an absence (Liu et al., 2011).”
Protected areas network is not adequate to protect a critically endangered East Africa Chelonian: Modelling distribution of pancake tortoise, Malacochersus tornieri under current and future climates (2020, methodology): “To perform many types of model evaluation, metrics of model fit are needed (cf. Liu et al. 2010).”
Selecting thresholds for the prediction of species occurrence with presence‐only data (2013, methodology): “The selected thresholds were applied to independent test data, and the accuracy of the binary predictions was evaluated with four accuracy measures (see Liu et al., 2011, for details): sensitivity (Se), specificity (Sp), true skill statistic (TSS) and kappa.”
Can species distribution modelling provide estimates of population densities? A case study with jaguars in the Neotropics (2012, methodology): “Second, the TSS was suggested as an alternative to Cohen’s kappa, which is also a common evaluation metric in SDM, but was proved to be sensitive to the species’ overall prevalence (Allouche et al., 2006; Liu et al., 2011).”
Global hotspots and correlates of emerging zoonotic diseases (2017, methodology): “For the ten-fold cross-validation runs, we calculated the AUC, a threshold-independent measure of model predictive performance that is commonly used as a validation metric in species distribution modelling 63 .”
Classification of the COVID-19 infected patients using DenseNet201 based deep transfer learning (2020, methodology): “A c can be computed as (Liu et al., 2011): However, A c does not consider how the data is distributed.”