预测物种潜在分布区——比较SVM与GARP

doi:10.17521/cjpe.2007.0091

摘要/Abstract

摘要：

物种分布与环境因子之间存在着紧密的联系,因此利用环境因子作为预测物种分布模型的变量是当前最普遍的建模思路,但是绝大多数物种分布预测模型都遇到了难以解决的“高维小样本"问题。该研究通过理论和实践证明,基于结构风险最小化原理的支持向量机(Support vector machine, SVM)算法非常适合“高维小样本"的分类问题。以20种杜鹃花属(Rhododendron)中国特有种为检验对象,利用标本数据和11个1 km×1 km的栅格环境数据层作为模型变量,预测其在中国的潜在分布区,并通过全面的模型评估——专家评估,受试者工作特征(Receiver operator characteristic, ROC)曲线和曲线下方面积(Area under the curve, AUC)——来比较模型的性能。我们实现了以SVM为核心的物种分布预测系统,并且通过试验证明其无论在计算速度还是预测效果上都远远优于当前广泛使用的规则集合预测的遗传算法(Algorithm for rule-set prediction, GARP)预测系统。

关键词: 物种分布预测模型, 支持向量机, GARP, ROC曲线, 杜鹃花属, 潜在分布区

Abstract:

Aims The most common method to build a predictive model of species' potential distribution is to use environmental factors, because they strongly affect species distribution. Unfortunately, most predictive models suffer from the “high dimension small sample size" problem, and cannot give satisfactory results in many cases. Support vector machine (SVM), which is based on structural risk minimization principle, has proven to be especially suitable for such data by both theory and abundant applications. Our objective was to implement a new predictive system of species' potential distribution based on the SVM method.
Methods We performed a country-scale case study using 20 Chinese endemic species of Rhododendron, employing herbarium specimen data and 11 layers of 1 km×1 km digital environmental grid data. Through expert evaluation and receiver operator characteristic (ROC) curve, we compared SVM predictions with those of a commonly used modeling method, the genetic algorithm for rule-set prediction (GARP).
Important findings All scores of SVM's prediction are higher than GARP's in expert evaluation. For the statistical analysis of ROC curve, almost all the area under the curve (AUC) determinations of SVM are larger than that of GARP. Furthermore, SVM's prediction speed is much faster than GARP's. Through our experiment, comprehensive evaluation proved that SVM is much better than GARP in terms of both performance and accuracy on the “high dimension small sample size" problem.

Key words: predictive model of species distribution, support vector machine (SVM), genetic algorithm for rule-set prediction (GARP), receiver operator characteristic (ROC) curve, Rhododendron, potential distribution

左闻韵, 劳逆, 耿玉英, 马克平. 预测物种潜在分布区——比较SVM与GARP. 植物生态学报, 2007, 31(4): 711-719. DOI: 10.17521/cjpe.2007.0091

ZUO Wen-Yun, LAO Ni, GENG Yu-Ying, MA Ke-Pin. PREDICTING SPECIES' POTENTIAL DISTRIBUTION—SVM COMPARED WITH GARP. Chinese Journal of Plant Ecology, 2007, 31(4): 711-719. DOI: 10.17521/cjpe.2007.0091

导出引用管理器 EndNote|Ris|BibTeX

链接本文: https://www.plant-ecology.com/CN/10.17521/cjpe.2007.0091

https://www.plant-ecology.com/CN/Y2007/V31/I4/711

图/表 7

参考文献 42

[1]	Anderson RP, Lew D, Peterson AT (2003). Evaluating predictive models of species' distributions: criteria for selecting optimal models. Ecological Modelling, 162, 211-232. DOI URL
[2]	Anderson RP, Martínez-Meyer E(2004). Modeling species' geographic distributions for preliminary conservation assessments: an implementation with the spiny pocket mice ( Heteromys) of Ecuador. Biological Conservation, 116, 167-179. DOI URL
[3]	Bolliger J, Kienast F, Bugmann H (2000). Comparing models for tree distributions: concept, structures, and behavior. Ecological Modelling, 134, 89-102. DOI URL
[4]	Brown JH, Lomolino MV (1998). Biogeography 2nd edn. Sinauer Associates, Sunderland, Massachusetts.
[5]	Busby JR (1986). A biogeographical analysis of Nothofagus cunninghamii (Hook.) Oerst. in southeastern Australia. Australian Journal of Ecology, 11, 1-7.
[6]	Carpenter G, Gillison AN, Winter J (1993). DOMAIN: a flexible modeling procedure for mapping potential distributions of plants, animals. Biodiversity and Conservation, 2, 667-680. DOI URL
[7]	Chang CC, Lin CJ (2006). LIBSVM: a library for support vector machines. Available via DIALOG. http://www.csie.ntu.edu.tw/~cjlin/papers/libsvm.pdf. Cited 15 Jan 2006.
[8]	Fang RZ (方瑞征), Min TL (闵天禄) (1995). The floristic study on the genus Rhododendron. Acta Botanica Yunnanica (云南植物研究), 17, 359-379. (in Chinese with English abstract)
[9]	Farber O, Kadmon R (2003). Assessment of alternative approaches for bioclimatic modeling with special emphasis on the Mahalanobis distance. Ecological Modelling, 160, 115-130. DOI URL
[10]	Gunn SR (1998). Support vector machines for classification and regression. Available via DIALOG. http://www.ecs.soton.ac.uk/~srg/publications. Cited 21 Jul 2006.
[11]	Guo Q, Kelly M, Graham CH (2005). Support vector machines for predicting distribution of sudden oak death in California. Ecological Modelling, 182, 75-90. DOI URL
[12]	Hirzel AH, Helfer V, Metral F (2001). Assessing habitat-suitability models with a virtual species. Ecological Modelling, 145, 111-121. DOI URL
[13]	Iverson LR, Prasad AM (1998). Predicting abundance of 80 tree species following climate change in the eastern United States. Ecological Monographs, 68, 465-485. DOI URL
[14]	Lin CJ (2006). LIBSVM. http://www.csie.ntu.edu.tw/~cjlin. Cited 21 Jul 2006.
[15]	Manel S, Dias JM, Ormerod SJ (1999). Comparing discriminant analysis, neural networks and logistic regression for predicting species distributions: a case study with a Himalayan river bird. Ecological Modelling, 120, 337-347. DOI URL
[16]	Manrique CE, Duran R, Argaez J (2003). Phytogeographic analysis of taxa endemic to the Yucatán Peninsula using geographic information systems, the domain heuristic method and parsimony analysis of endemicity. Diversity and Distributions, 9, 313-330. DOI URL
[17]	Martínez-Meyer E, Peterson AT, Hargrove WW (2004). Ecological niches as stable distributional constraints on mammal species, with implications for Pleistocene extinctions and climate change projections for biodiversity. Global Ecology and Biogeography, 13, 305-314. DOI URL
[18]	Mozer MC, Dodier R, Colagrosso MD, Guerra-Salcedo C, Wolniewicz R 2002. Prodding the ROC curve: constrained optimization of classifier performance. In: Dietterich T, Becker S, Ghahramani Z eds. Advances in Neural Information Processing Systems XIV. MIT Press, Cambridge, MA, 1409-1415.
[19]	Nix HA 1986. A biogeographic analysis of Australian elapid snakes. In: Longmore R ed. Atlas of Elapid Snakes of Australia. Australian Government Publishing Service, Canberra, 4-15.
[20]	Osborne PE, Alonso JC, Bryant RG (2001). Modelling landscape-scale habitat use using GIS and remote sensing: a case study with great bustards. Journal of Applied Ecology, 38, 458-471. DOI URL
[21]	Ozesmi SL, Ozesmi U (1999). An artificial neural network approach to spatial habitat modelling with interspecific interaction. Ecological Modelling, 116, 15-31. DOI URL
[22]	Peterson AT (2001). Predicting species' geographic distributions based on ecological niche modeling. Condor, 103, 599-605. DOI URL
[23]	Peterson AT, Cohoon KP (1999). Sensitivity of distributional prediction algorithms to geographic data completeness. Ecological Modelling, 117, 159-164. DOI URL
[24]	Peterson AT, Ortega-Huerta MA, Bartley J, Sanchez-Cordero V, Soberonk J, Buddemeier RH, Stockwell DRB (2002). Future projections for Mexican faunas under global climate change scenarios. Nature, 416, 626-629. DOI URL PMID
[25]	Peterson AT, Robins CR (2003). Using ecological-niche modeling to predict barred owl invasions with implications for spotted owl conservation. Conservation Biology, 17, 1161-1165. DOI URL
[26]	Peterson AT, Sanchez-Cordero V, Soberon J, Bartley J, Buddemeier RW, Navarro-Siguenza AG (2001). Effects of global climate change on geographic distributions of Mexican Cracidae. Ecological Modelling, 144, 21-30. DOI URL
[27]	Peterson AT, Soberon J, Sanchez-Cordero V (1999). Conservatism of ecological niches in evolutionary time. Science, 285, 1265-1267. DOI URL PMID
[28]	Phillips SJ, Anderson RP, Schapire RE (2006). Maximum entropy modeling of species geographic distributions. Ecological Modelling, 190, 231-259. DOI URL
[29]	Robertson MP, Caithness N, Villet MH (2001). A PCA-based modelling technique for predicting environmental suitability for organisms from presence records. Diversity and Distributions, 7, 15-27. DOI URL
[30]	Stockwell D, Peters D (1999). The GARP modelling system: problems and solutions to automated spatial prediction. International Journal of Geographical Information Science, 13, 143-158. DOI URL
[31]	Stockwell D, Peterson AT (2003). Comparison of resolution of methods used in mapping biodiversity patterns from point-occurrence data. Ecological Indicators, 3, 213-221. DOI URL
[32]	Stockwell DRB (1997). Generic predictive systems: an empirical evaluation using the learning base system (LBS). Expert Systems with Applications, 12, 301-310. DOI URL
[33]	Stockwell DRB (2006). Improving ecological niche models by data mining large environmental datasets for surrogate models. Ecological Modelling, 193, 188-196.
[34]	Stockwell DRB, Beach JH, Stewart A, Vorontsov G, Vieglais D, Pereira RS (2006). The use of the GARP genetic algorithm and Internet grid computing in the Lifemapper world atlas of species biodiversity. Ecological Modelling, 195, 139-145. DOI URL
[35]	Stockwell DRB, Noble IR (1992). Induction of sets of rules from animal distribution data: a robust and informative method of data analysis. Mathematics and Computers in Simulation, 33, 385-390. DOI URL
[36]	Stockwell DRB, Peterson AT (2002). Effects of sample size on accuracy of species distribution models. Ecological Modelling, 148, 1-13. DOI URL
[37]	Sutherst RW, Maywald GF (1985). A computerised system for matching climates in ecology. Agriculture, Ecosystems & Environment, 13, 281-299.
[38]	Tellez-Valdes O, DaVila-Aranda P (2003). Protected areas and climate change: a case study of the cacti in the Tehuacan-Cuicatlan Biosphere Reserve, Mexico. Conservation Biology, 17, 846-853. DOI URL
[39]	The University of Kansas Center for Research (2002). User's manual. http://www.lifemapper.org/desktopgarp/Default.asp?Item=1&Lang=1. Cited 21 Jul 2006.
[40]	Vapnik VN (1995). The Nature of Statistical Learning Theory. Springer-Verlag, Berlin.
[41]	Wu ZY, Peter HR, Hong DY (2005). Flora of China 14. Science Press, Beijing, China,
[42]	Yu GR (于贵瑞), He HL (何洪林), Liu XA (刘新安), Liu D (刘栋), Wang QF (王秋凤), Ren CY (任传友), Li ZQ (李正泉), Su W (苏文), Yue YZ (岳燕珍), Fan LS (范辽生), Guo XB (郭学兵), Zhu QL (祝青林), Cai F (蔡福) (2004). Atlas for Spatialized Information of Terrestrial Ecosystem in China—Volume of Climatological Elements (中国陆地生态系统空间化信息研究图集——气候要素分卷). China Meteorological Press, Beijing,

	实际真值Actual positive	实际假值Actual negative
预测真值Predicted positive	真真True positive (TP)	假真False positive (FP)
预测假值Predicted negative	假假False negative (FN)	真假True negative (TN)

	实际真值Actual positive	实际假值Actual negative
预测真值Predicted positive	真真True positive (TP)	假真False positive (FP)
预测假值Predicted negative	假假False negative (FN)	真假True negative (TN)

	GARP	SVM
雪山杜鹃Rhododendron aganniphum	0	3
短花杜鹃Rhododendron brachyanthum	0.5	3
美容杜鹃Rhododendron calophytum	0.5	3
腺果杜鹃Rhododendron davidii	0.5	3
大白杜鹃Rhododendron decorum	0	2
树生杜鹃Rhododendron dendrocharis	0.5	2.5
似血杜鹃Rhododendron haematodes	0.5	0.5
亮鳞杜鹃Rhododendron heliolepis	0	2
露珠杜鹃Rhododendron irroratum	0.5	2.5
黄花杜鹃Rhododendron lutescens	0	3
雪层杜鹃Rhododendron nivale	0	3.5
马银花Rhododendron ovatum	0	5
栎叶杜鹃Rhododendron phaeochrysum	0	3.5
大树杜鹃Rhododendron protistum var. giganteum	0	1.5
血红杜鹃Rhododendron sanguineum	0	1.5
多变杜鹃Rhododendron selense	0	0
杜鹃Rhododendron simsii	0	5
芒刺杜鹃Rhododendron strigillosum	0	1.5
紫玉盘杜鹃Rhododendron uvarifolium	0	2.5
黄杯杜鹃Rhododendron wardii	0	1.5

	GARP	SVM
雪山杜鹃Rhododendron aganniphum	0	3
短花杜鹃Rhododendron brachyanthum	0.5	3
美容杜鹃Rhododendron calophytum	0.5	3
腺果杜鹃Rhododendron davidii	0.5	3
大白杜鹃Rhododendron decorum	0	2
树生杜鹃Rhododendron dendrocharis	0.5	2.5
似血杜鹃Rhododendron haematodes	0.5	0.5
亮鳞杜鹃Rhododendron heliolepis	0	2
露珠杜鹃Rhododendron irroratum	0.5	2.5
黄花杜鹃Rhododendron lutescens	0	3
雪层杜鹃Rhododendron nivale	0	3.5
马银花Rhododendron ovatum	0	5
栎叶杜鹃Rhododendron phaeochrysum	0	3.5
大树杜鹃Rhododendron protistum var. giganteum	0	1.5
血红杜鹃Rhododendron sanguineum	0	1.5
多变杜鹃Rhododendron selense	0	0
杜鹃Rhododendron simsii	0	5
芒刺杜鹃Rhododendron strigillosum	0	1.5
紫玉盘杜鹃Rhododendron uvarifolium	0	2.5
黄杯杜鹃Rhododendron wardii	0	1.5