Maxent (software)
| Maxent | |
|---|---|
| Basic data
|
|
| developer | S. Phillips, M. Dudik et al. R. Schapire, supported by AT&T Labs-Research, Princeton University and Center for Biodiversity and Conservation, American Museum of Natural History |
| Current version | 3.4.1 (April 2017) |
| operating system | platform independent ( Java ) |
| category | Application program |
| License | MIT license |
| German speaking | No |
| biodiversityinformatics.amnh.org | |
Maxent is a freely available software that is used to predict the potential distribution ( habitat modeling ) of animal or plant species on the basis of point-by-point distribution information and certain environmental factors . Maxent has been open source since 2017 .
background
Knowing about the distribution of certain species and understanding their demands on the habitat is important, for example, if suitable protected areas are to be created. The delimitation of the suitable area on the basis of often only a few isolated information about the location is difficult. That is why modern ecology is increasingly making use of biodiversity informatics to solve such questions . Maxent is here in a number of different software solutions for distribution modeling (such as BIOCLIM, GARP and others).
Maxent was initially programmed in Java. An implementation for the statistics software R (package maxnet , based on logistic regression ) is available. B. graphical representation options are missing.
Working method
Modeling by Maxent is based on the maximum entropy method , a machine learning technique . Maxent needs geo -referenced data as input variables : On the one hand, distribution information as point information (X and Y coordinates) of the animal or plant species to be modeled, on the other hand, suitable environmental variables (e.g. climatic factors such as precipitation or temperature, information on altitude, topography, etc.) in the ASCII -based .asc format from ESRI . As a result, the software delivers a map that shows the probability of the potential occurrence of the species in a certain area, as well as additional statistical interpretation aids, such as an indication of the relative participation of variables in the modeling, an ROC curve for evaluating the model quality, impact curves . : response curves) of the entered environmental variables, which show the dependence of the model quality on the characteristics of the respective environmental variables, as well as a jackknife test , which shows the portion of the information added to the model for each environmental variable. The data preparation and visualization of the result is usually done with the help of geographic information systems .
literature
- Steven J. Phillips, Robert P. Anderson, Robert E. Schapire: A Maximum Entropy Approach to Species Distribution Modeling . Proceedings of the twenty-first international conference on Machine learning, Banff, Alberta, Canada, 2004, p. 83 ff. Research.att.com (PDF)
- Steven J. Phillips, Miroslav Dudík, Robert E. Schapire: Maximum entropy modeling of species geographic distributions . In: Ecological Modeling , 190, 2006, pp. 231-259. cs.princeton.edu (PDF)
- Steven J. Phillips, Robert P. Anderson, Miroslav Dudík, Robert E. Schapire, Mary E. Blair (2017): Opening the black box: an open-source release of Maxent. In: Ecography , 40.7, pp. 887-893, doi: 10.1111 / ecog.03049
Web links
- Maxent Homepage (English)
Individual evidence
- ↑ Jane Elith et al .: Novel methods improve prediction of species' distributions from occurrence data . In: Ecography , 29, 2006, pp. 129-151, doi: 10.1111 / j.2006.0906-7590.04596.x .
- ↑ Steven J. Phillips: A Brief Tutorial on Maxent . ( Memento of the original from June 17, 2010 in the Internet Archive ) Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. ( MS Word ; 1.5 MB)
