Opening the black box: an open-source release of Maxent

doi:10.1111/ECOG.03049

Home
/
Papers
/
Opening the black box: an open-source release of Maxent

Journal Article•DOI•

Opening the black box: an open-source release of Maxent

Steven J. Phillips¹, Robert P. Anderson², Robert P. Anderson¹, Miroslav Dudík³, Robert E. Schapire³, Mary E. Blair¹ - Show less +2 more•Institutions (3)

American Museum of Natural History¹, City University of New York², Microsoft³

01 Jul 2017-Ecography (Blackwell Publishing Ltd)-Vol. 40, Iss: 7, pp 887-893

TL;DR: A new open-source release of the Maxent software for modeling species distributions from occurrence records and environmental data is announced, and a new R package for fitting Maxent models using the glmnet package for regularized generalized linear models is described.

read less

Abstract: This software note announces a new open-source release of the Maxent software for modeling species distributions from occurrence records and environmental data, and describes a new R package for fitting such models. The new release (ver. 3.4.0) will be hosted online by the American Museum of Natural History, along with future versions. It contains small functional changes, most notably use of a complementary log-log (cloglog) transform to produce an estimate of occurrence probability. The cloglog transform derives from the recently-published interpretation of Maxent as an inhomogeneous Poisson process (IPP), giving it a stronger theoretical justification than the logistic transform which it replaces by default. In addition, the new R package, maxnet, fits Maxent models using the glmnet package for regularized generalized linear models. We discuss the implications of the IPP formulation in terms of model inputs and outputs, treating occurrence records as points rather than grid cells and interpreting the exponential Maxent model (raw output) as as an estimate of relative abundance. With these two open-source developments, we invite others to freely use and contribute to the software.

...read moreread less

Citations

PDF

Open Access

More filters

Journal Article•DOI•

Statistics for Spatial Data.

[...]

Andrew B. Lawson¹, Noel A Cressie•Institutions (1)

University of Dundee¹

01 Mar 1993-The Statistician

6,278 citations

Journal Article•DOI•

Identification of novel bat coronaviruses sheds light on the evolutionary origins of SARS-CoV-2 and related viruses.

[...]

Hong Zhou¹, Jingkai Ji¹, Xing Chen², Yuhai Bi³, Juan Li¹, Qihui Wang³, Tao Hu¹, Hao Song³, Runchu Zhao³, Yanhua Chen², Mingxue Cui¹, Yanyan Zhang¹, Alice C. Hughes³, Alice C. Hughes², Edward C. Holmes⁴, Weifeng Shi¹ - Show less +12 more•Institutions (4)

Peking Union Medical College¹, Xishuangbanna Tropical Botanical Garden², Chinese Academy of Sciences³, University of Sydney⁴

19 Aug 2021-Cell

TL;DR: A meta-transcriptomic study of 411 bat samples collected from a small geographical region in Yunnan province, China, between May 2019 and November 2020 as mentioned in this paper identified 24 full-length coronavirus genomes.

...read moreread less

210 citations

Journal Article•DOI•

ENMeval 2.0: Redesigned for customizable and reproducible modeling of species’ niches and distributions

[...]

Jamie M. Kass¹, Jamie M. Kass², Robert Muscarella³, Peter J. Galante⁴, Corentin L. Bohl, Gonzalo E. Pinilla-Buitrago¹, Robert A. Boria⁵, Mariano Soley-Guardia⁶, Robert P. Anderson¹, Robert P. Anderson⁴ - Show less +6 more•Institutions (6)

City University of New York¹, Okinawa Institute of Science and Technology², Uppsala University³, American Museum of Natural History⁴, University of California, Merced⁵, University of Costa Rica⁶

01 Sep 2021-Methods in Ecology and Evolution

TL;DR: An insufficient reporting of model performance and parameterization, heavy reliance on model selection with AICc and low utilization of spatial cross‐validation are found; it is explained how ENMeval 2.0 can help address these issues.

...read moreread less

158 citations

Cites methods from "Opening the black box: an open-sour..."

...Opening the black box: An open- source release of Maxent....
[...]
...The authors thank Steven J. Phillips for development of the clamping function, the Maxent and Wallace Google Group users for invaluable input, Mary E. Blair, the Anderson Lab group, and two anonymous reviewers for constructive comments on this manuscript....
[...]
...ENMeval was the first R package to make such evaluations (often termed model tuning) widely accessible for the Maxent algorithm....
[...]
...When users prefer to select models with AICc, which comes with a statistical caveat for Maxent (Warren & Seifert, 2011), ENMeval 2.0 can additionally quantify how well such models predict withheld data compared with null models (Kass et al., 2020)....
[...]
...The algorithms Maxent (Phillips et al., 2017), via maxnet (R package) and maxent.jar (Java) and BIOCLIM (with R package dismo; Hijmans et al., 2020) are now implemented as example ENMdetails objects to illustrate the capacity for expansion (details for BIOCLIM implementation in Appendix 2)....
[...]

Journal Article•DOI•

Social physics

[...]

Fatemeh Jafarlou¹•Institutions (1)

Complexity Science Hub Vienna¹

01 Feb 2022-Physics Reports

TL;DR: The field of social physics has been a hot topic in the last few decades as mentioned in this paper , with many researchers venturing outside of their traditional domains of interest, but also taking from physics the methods that have proven so successful throughout the 19th and the 20th century.

...read moreread less

133 citations

Journal Article•DOI•

Using species distribution models at local scale to guide the search of poorly known species: Review, methodological issues and future directions

[...]

Mauro Fois¹, Alba Cuena-Lombraña¹, Alba Cuena-Lombraña², Giuseppe Fenu¹, Gianluigi Bacchetta¹ - Show less +1 more•Institutions (2)

University of Cagliari¹, Sapienza University of Rome²

10 Oct 2018-Ecological Modelling

TL;DR: Among specific applications of species distribution models (SDMs), the use of SDMs probabilistic maps for guiding field surveys is increasingly applied, but the efficiency of different algorithms, metrics for model evaluation and algorithm-specific settings have not yet been sufficiently investigated.

...read moreread less

131 citations

Cites methods from "Opening the black box: an open-sour..."

...In the absence of adequate variance partitioning methods for MaxEnt, permutation importance values were used as gross estimates of the variance explained by variables or groups of variables in SDM....
[...]
...A more recent version of MaxEnt (ver. 3.4.1; Phillips et al., 2017) is currently available; nonetheless, we preferred using the previous version since it is still the most tested one and because no comparisons with concrete examples, highlighting differences between the two versions, are nowadays available....
[...]
...Among modelling techniques, the presence-only or presence-background methods were the most frequently applied with MaxEnt (15 studies) and GARP (six studies), while, among presence-absence methods, ENFA was the most fre UN CO RR EC TE D PR OO F Table 1 Summary, in chronological order, of the 28 studies applying species distribution models (SDM) to guide searches for new populations....
[...]
...The variable importance was measured according to the permutation importance, which is calculated by randomly permuting training presence and pseudo-absence data in MaxEnt....
[...]
...UN CO RR EC TE D PR OO F Ecological Modelling xxx (2018) xxx-xxx Contents lists available at ScienceDirect Ecological Modelling journal homepage: www.elsevier.com Using species distribution models at local scale to guide the search of poorly known species: Review, methodological issues and future directions Mauro Foisa, ⁎, Alba Cuena-Lombrañab, Giuseppe Fenua, Gianluigi Bacchetta a, c a Centro Conservazione Biodiversità, Dipartimento di Scienze della Vita e dell’Ambiente, Università degli Studi di Cagliari, Viale S. Ignazio da Laconi, 13, Cagliari 09123, Italy b Dipartimento di Biologia Ambientale, ‘Sapienza’ Università di Roma, P.le A. Moro 5, 00185 Roma, Italy c Hortus Botanicus Karalitanus (HBK), Università degli Studi di Cagliari, Viale Sant’Ignazio da Laconi, 9–11, Cagliari 09123, Italy A R T I C L E I N F O Keywords: Ground validation MaxEnt Mediterranean flora Independent presence-absence data Plant distribution patterns Regularization multiplier A B S T R A C T Among specific applications of species distribution models (SDMs), the use of SDMs probabilistic maps for guiding field surveys is increasingly applied....
[...]

1
2
3
4
…
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200

Collapse

References

PDF

Open Access

More filters

Journal Article•DOI•

Regularization Paths for Generalized Linear Models via Coordinate Descent

[...]

Jerome H. Friedman¹, Trevor Hastie¹, Robert Tibshirani•Institutions (1)

Stanford University¹

02 Feb 2010-Journal of Statistical Software

TL;DR: In comparative timings, the new algorithms are considerably faster than competing methods and can handle large problems and can also deal efficiently with sparse features.

...read moreread less

Abstract: We develop fast algorithms for estimation of generalized linear models with convex penalties. The models include linear regression, two-class logistic regression, and multinomial regression problems while the penalties include l(1) (the lasso), l(2) (ridge regression) and mixtures of the two (the elastic net). The algorithms use cyclical coordinate descent, computed along a regularization path. The methods can handle large problems and can also deal efficiently with sparse features. In comparative timings we find that the new algorithms are considerably faster than competing methods.

...read moreread less

13,656 citations

Journal Article•DOI•

Maximum entropy modeling of species geographic distributions

[...]

Steven J. Phillips¹, Robert P. Anderson², Robert P. Anderson³, Robert E. Schapire⁴•Institutions (4)

AT&T Labs¹, City University of New York², American Museum of Natural History³, Princeton University⁴

25 Jan 2006-Ecological Modelling

TL;DR: In this paper, the use of the maximum entropy method (Maxent) for modeling species geographic distributions with presence-only data was introduced, which is a general-purpose machine learning method with a simple and precise mathematical formulation.

...read moreread less

13,120 citations

Journal Article•DOI•

Novel methods improve prediction of species' distributions from occurrence data

[...]

Jane Elith¹, Catherine H. Graham², Robert P. Anderson³, Miroslav Dudík⁴, Simon Ferrier, Antoine Guisan⁵, Robert J. Hijmans⁶, Falk Huettmann⁷, John R. Leathwick⁸, Anthony Lehmann, Jin Li⁹, Lúcia G. Lohmann¹⁰, Bette A. Loiselle¹¹, Glenn Manion, Craig Moritz⁶, Miguel Nakamura¹², Yoshinori Nakazawa¹³, Jacob C. M. Mc Overton¹⁴, A. Townsend Peterson¹³, Steven J. Phillips¹⁵, Karen Richardson¹⁶, Ricardo Scachetti-Pereira, Robert E. Schapire, Jorge Soberón¹³, Stephen E. Williams¹⁷, Mary S. Wisz, Niklaus E. Zimmermann¹⁸ - Show less +23 more•Institutions (18)

University of Melbourne¹, Stony Brook University², City University of New York³, Princeton University⁴, University of Lausanne⁵, University of California, Berkeley⁶, University of Alaska Fairbanks⁷, National Institute of Water and Atmospheric Research⁸, Commonwealth Scientific and Industrial Research Organisation⁹, University of São Paulo¹⁰, University of Missouri¹¹, Consejo Nacional de Ciencia y Tecnología¹², University of Kansas¹³, Landcare Research¹⁴, AT&T¹⁵, McGill University¹⁶, James Cook University¹⁷, Swiss Federal Institute for Forest, Snow and Landscape Research¹⁸

01 Apr 2006-Ecography

TL;DR: This work compared 16 modelling methods over 226 species from 6 regions of the world, creating the most comprehensive set of model comparisons to date and found that presence-only data were effective for modelling species' distributions for many species and regions.

...read moreread less

Abstract: Prediction of species' distributions is central to diverse applications in ecology, evolution and conservation science. There is increasing electronic access to vast sets of occurrence records in museums and herbaria, yet little effective guidance on how best to use this information in the context of numerous approaches for modelling distributions. To meet this need, we compared 16 modelling methods over 226 species from 6 regions of the world, creating the most comprehensive set of model comparisons to date. We used presence-only data to fit models, and independent presence-absence data to evaluate the predictions. Along with well-established modelling methods such as generalised additive models and GARP and BIOCLIM, we explored methods that either have been developed recently or have rarely been applied to modelling species' distributions. These include machine-learning methods and community models, both of which have features that may make them particularly well suited to noisy or sparse information, as is typical of species' occurrence data. Presence-only data were effective for modelling species' distributions for many species and regions. The novel methods consistently outperformed more established methods. The results of our analysis are promising for the use of data from museums and herbaria, especially as methods suited to the noise inherent in such data improve.

...read moreread less

7,589 citations

"Opening the black box: an open-sour..." refers methods in this paper

...When run on the data set of Elith et al. (2006), maxnet has similar performance to the Maxent Java application....
[...]

Journal Article•DOI•

Statistics for Spatial Data.

[...]

Andrew B. Lawson¹, Noel A Cressie•Institutions (1)

University of Dundee¹

01 Mar 1993-The Statistician

6,278 citations

"Opening the black box: an open-sour..." refers methods in this paper

...An IPP is a widely-used model for a random set Z of points falling in some domain D (Cressie 1993, Diggle 2003)....
[...]
...Because Maxent is an IPP, standard generalized linear modeling software can be used to fit Maxent models via Poisson regression (Renner and Warton 2013), or even more conveniently, using standard logistic regression (Fithian and Hastie 2013). Specifically, the latter authors showed that the coefficients b of the Maxent or IPP model can be fitted via a weighting process they call infinitely-weighted logistic regression (IWLR). The idea is to fit a logistic model to occurrence records (with response variable y  1) and background data (points chosen randomly from the domain D, with response variable y  0). This process yields coefficients b for an exponential model, and has been used in studies of resource selection by animals (Manly et al. 2002), but may not produce the same values of the coefficients as Maxent. The novel contribution of Fithian and Hastie (2013) was to give a large weight W to all the background data and to show that the limit (as W tends to infinity) of the resulting vector of logistic regression coefficients equals the Maxent (and IPP) coefficients....
[...]

Journal Article•DOI•

5. Statistics for Spatial Data

[...]

Mike Rees¹, N. Cressie•Institutions (1)

University of Westminster¹

01 Jan 1993-Journal of The Royal Statistical Society Series A-statistics in Society

TL;DR: Cressie et al. as discussed by the authors presented the Statistics for Spatial Data (SDS) for the first time in 1991, and used it for the purpose of statistical analysis of spatial data.

...read moreread less

Abstract: 5. Statistics for Spatial Data. By N. Cressie. ISBN 0 471 84336 9. Wiley, Chichester, 1991. 900 pp. £71.00.

...read moreread less

5,555 citations