A maximum entropy approach to species distribution modeling

By Steven J. Phillips, Miroslav Dud'ik, Robert E. Schapire. International Conference on Machine Learning, 2004

2,411 citations recorded by the source corpus.

Subfields: Ecological Modeling, Ecology, Genetics

View Paper: https://www.semanticscholar.org/paper/c63546ae2b47eb18bdfda8a50178bffeb24b494e

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 Genetics

Citations in the corpus

Multi-model ensemble projections of climate change effects on global marine biodiversity (2015, background): “These are Maxent (Phillips et al., 2004; Phillips, 2008), AquaMaps (Kaschner et al.”
Global mapping of Jatropha curcas yield based on response of fitness to present and future climate (2010, methodology): “This geographic distribution defines where environmental conditions are more suitable for species occurrence, including areas that have not been naturally ‘colonized’, due to biotic competition or geographic barriers (Hutchinson, 1957; Phillips et al., 2004).”
GIS‐based niche models identify environmental correlates sustaining a contact zone between three species of European vipers (2008, methodology): “However, the issue of selecting an optimized threshold with Maxent and selecting the ‘best’ model from a model ensemble is rather subjective (Phillips et al., 2004; Phillips et al., 2006).”
Two alternative evaluation metrics to replace the true skill statistic in the assessment of species distribution models (2019, methodology): “In this context, it has to be stressed that the default number of background points (10,000) in MAXENT, a broadly adopted maximum entropy machine-learning SDM algorithm (Phillips et al. 2004, 2017), is often insufficient and – unless increased sufficiently to capture the range of existing environmen”
Use of Maximum Entropy Modeling in Wildlife Research (2009, background): “This approach estimates the most uniform distribution (maximum entropy) of sampling points compared to background locations given the constraints derived from the data [5,7,8].”
Predicting species distributions from museum and herbarium records using multiresponse models fitted with multivariate adaptive regression splines (2007, background): “E-mail: j.elith@unimelb.edu.au © 2007 The Authors 266 Diversity and Distributions , 13 , 265–275, Journal compilation © 2007 Blackwell Publishing Ltd urgent need to manage and conserve biodiversity, there has been an increasing focus on their use in conservation planning — not only for estimating sp”
A strict maximum likelihood explanation of MaxEnt, and some implications for distribution modelling (2013, methodology): “Raw output (q) The term ʻraw outputʼ, which is inherited from the Maxent software implementation of MaxEnt (Phillips et al. 2004, 2006, Phillips & Dudík 2008, Elith et al. 2011), is used for the model estimates q(Xl) as such.”
Modelling climate-change-induced shifts in the distribution of the koala (2011, methodology): “We chose Maxent for our modelling because it can use presence only data, can utilise both continuous and categorical data (Phillips et al. 2004) and has been found to be a robust modelling technique (e.”
Modelling invasion for a habitat generalist and a specialist plant species (2008, methodology): “Alternatively, models are generally designed to identify ecosystem characteristics that can predict habitat suitability on the landscape, and broadly applied to different species (Phillips et al ., 2004; Payne & Stockwell, 2006).”
A Survey on ensemble learning under the era of deep learning (2021, methodology): “…the landslide susceptibility mapping, Chen et al. (Chen et al. 2017) proposed several ensembles using multiple base learners that were heterogeneously trained with artificial neural network (ANN) (Rumelhart et al. 1986), maximum entropy (MaxEnt) (Phillips et al. 2004) and SVM (Chang and Lin 2011).”
Maxent modeling for predicting suitable habitat for threatened and endangered tree Canacomyrica monticola in New Caledonia (2009, methodology): “Since our sample size was low (< 20) we used only linear and quadratic features and maintained other settings as default (Phillips et al., 2004).”