How to assess the prediction accuracy of species presence-absence models without absence data?

doi:10.1111/j.1600-0587.2013.07585.x

	How to assess the prediction accuracy of species presence-absence models without absence data?
	Li, Wenkai 1; Guo, Qinghua1
	2013
发表期刊	ECOGRAPHY
ISSN	0906-7590
卷号	36 期号:7 页码:788-799
摘要	It is very common that only presence data are available in ecological niche modeling. However, most existing methods for evaluating the accuracy of presence-absence (binary) predictions of species require presence-absence data. The aim of this study is to present a new method for accuracy assessment that does not rely on absence data. Two new statistics Fpb and Fcpb were derived based on presence-background data. With generated six virtual species, we used DOMAIN, generalized linear modeling (GLM), and maximum entropy (MAXENT) to produce different species presence-absence predictions. To investigate the effectiveness of the new statistics in accuracy assessment, we used Fpb, Fcpb, the traditional F-measure (F), kappa coefficient, true skill statistic (TSS), area under the receiver operating characteristic curve (AUC), and the contrast validation index (CVI) to evaluate the accuracy of predictions, and the behaviors of these accuracy measures were compared. The effectiveness of Fpb for threshold selection and estimation of species prevalence was also investigated. Experimental results show that Fcpb is an estimate of F. The Pearson's correlation coefficient (COR) between Fcpb and F is 0.9882, with a root-mean-square error (RMSE) of 0.0171. In general, Fpb, Fcpb, F, kappa coefficient, TSS, and CVI can sort models by the accuracy of binary prediction, but AUC is not appropriate to evaluate the accuracy of binary prediction. For DOMAIN, GLM, and MAXENT, finding the threshold by maximizing Fpb and by maximizing F result in similar accuracies. In addition, the estimation of species prevalence based on binary output with maximizing Fpb as the thresholding method is significantly more accurate than simply averaging the original continuous output. The best estimate of prevalence is provided by the binary output of MAXENT, with an RMSE of 0.0116. Finally, we conclude that the new method is promising in accuracy assessment, threshold selection, and estimation of species prevalence, all of which are important but challenging problems with presence-only data. Because it does not require absence data, the new method will have important applications in ecological niche modeling.
学科领域	Biodiversity Conservation ; Ecology
DOI	10.1111/j.1600-0587.2013.07585.x
收录类别	SCI
语种	英语
WOS关键词	PRESENCE-ONLY DATA ; HABITAT-SUITABILITY MODELS ; F-SCORE ; DISTRIBUTIONS ; SELECTION ; PERFORMANCE ; CALIFORNIA ; CLIMATE ; PLANTS ; ERRORS
WOS研究方向	Science Citation Index Expanded (SCI-EXPANDED)
WOS记录号	WOS:000320468900005
出版者	WILEY
文献子类	Article
出版地	HOBOKEN
EISSN	1600-0587
资助机构	National Science Foundation(National Science Foundation (NSF)) ; USGS National Climate Change and Wildlife Science Center ; National Science Foundation of China(National Natural Science Foundation of China (NSFC))
作者邮箱	Li, Wenkai/AFM-7916-2022 ; Guo, Qinghua/B-7731-2012
引用统计	被引频次：59[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://ir.ibcas.ac.cn/handle/2S10CLM1/27908
专题	植被与环境变化国家重点实验室
作者单位	1.Chinese Acad Sci, Inst Bot, State Key Lab Vegetat & Environm Change, Beijing, Peoples R China 2.Univ Calif Merced, Sch Engn, Merced, CA 95343 USA
推荐引用方式 GB/T 7714	Li, Wenkai,Guo, Qinghua. How to assess the prediction accuracy of species presence-absence models without absence data?[J]. ECOGRAPHY,2013,36(7):788-799.
APA	Li, Wenkai,&Guo, Qinghua.(2013).How to assess the prediction accuracy of species presence-absence models without absence data?.ECOGRAPHY,36(7),788-799.
MLA	Li, Wenkai,et al."How to assess the prediction accuracy of species presence-absence models without absence data?".ECOGRAPHY 36.7(2013):788-799.

条目包含的文件
文件名称/大小	文献类型	版本类型	开放类型	使用许可
How_to_assess_the_pr（1061KB）	期刊论文	出版稿	开放获取	CC BY-NC-SA	浏览请求全文