刘笑笑,王亮,徐胜华,梁勇

• 1:山东农业大学信息科学与工程学院
• 2:中国测绘科学研究院

摘要(Abstract)：

森林生物量的估算对于全球碳平衡和环境保护至关重要。通过遥感等手段提取与森林生物量相关的单波段特征、植被指数、纹理特征、地形因子等特征参数,特征数量往往较多,影响预测精度。该文提出了一种后向迭代的随机森林(RF-RFE)特征选择方法,即利用随机森林算法计算特征重要度,采用后向迭代的方法逐步简化特征参数。以内蒙古大兴安岭地区的激流河林场为研究区域,以实验区2012年"资源三号"遥感影像和森林资源3类调查的样地数据为数据源,使用RF-RFE算法进行特征选择分析。实验结果表明,在森林生物量遥感反演过程中的RF-RFE特征选择不但能降低时间复杂度,而且保证了特征选择的精度。

关键词(KeyWords)： 森林生物量;随机森林;后向迭代(RFE);特征选择

Abstract:

Keywords:

基金项目(Foundation): 国家“863计划”项目(2013AA122003);; 国家质量监督检验检疫总局公益性行业科研专项(201410308)

作者(Author): 刘笑笑,王亮,徐胜华,梁勇

DOI: 10.16251/j.cnki.1009-2307.2017.05.017

参考文献(References)：

• [1]刘琼阁,彭道黎,涂云燕.基于偏最小二乘的森林生物量遥感估测[J].东北林业大学学报,2014,42(7):44-47.(LIU Qiongge,PENG Daoli,TU Yunyan.Estimating forest biomass by partial least squares regression[J].Journal of Northeast University,2014,42(7):44-47.)
• [2]DESTRERO A,MOSCI S,MOL C,et al.Feature selection for high-dimensional data[J].Computer Manage Sci,2009(6):25-40.
• [3]PENG H C,LONG R,DING C.Feature selection based on mutual information criteria of max-dependency,max-relevance,and min-redundancy[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2005,27(8):1226-1238.
• [4]CHEN Xuewen.Margin-based wrapper methods for gene identification using microarray[J].Neurocomputing,2006,69(16-18):2236-2243..
• [5]MENZE B H,PETRICH W,HAMPRECHT F A.Multivariate feature selection and hierarchicalclassification for infrared spectroscopy:serum-based detection of bovine spongiform encephalopathy[J].Analytical and Bioanalytical Chemistry,2007,387(5):1801-1807.
• [6]张洪强,刘光远,赖祥伟.随机森林算法在肌电的重要特征选择中的应用[J].计算机科学,2013,40(1):200-202.(ZHANG Hongqiang,LIU Guangyuan,LAI Xiangwei.Application of random forest algorithm in improtant feature selection from EMG signal[J].Computer Science,2013,40(1):200-202.)
• [7]FANG J Y,CHEN A P,PENG CH,et al Changes in forest biomass carbon storage in China between 1949and 1998[J].Science,2001,292:2320-2322.
• [8]孟晓清,刘琪璟,陶立超,等.大兴安岭地区南北部幼中龄林碳储量研究[J].中南林业科技大学学报,2014,34(6):37-43.(MENG Xiaoqing,LIU Qijing,TAO Lichao,et al.Carbon storage of young and middle age forest in south and north Daxing'anling mountains[J].Journal of Central South University of Forestry&Technology,2014,34(6):37-43.)
• [9]DESTRERO A,SARKER L R,NICHOL J E.Improved forest biomass estimates using ALOS AVNIR-2texture indices[J].Remote Sensing of Environment,2011,115(4):968-977.
• [10]GRANITT P M,FURLANELLO C,BIASIOLI,et al.Recursive feature elimination with random forest for PTR-MS analysis of agroindustrial products[J].Chemometrics and Intelligent Laboratory Systems,2006(83):83-90.
• [11]PICARD R,HEALEY J.Digital Processing of Affective Signals[R].M.I.T Media Laboratory Perceptual Computing Section Technical Report,Massachusetts:MIT,1998:444.
• [12]李欣海.随机森林模型在分类与回归分析中的应用[J].应用昆虫学报,2013,50(4):1190-1197.(LI Xinhai.Using“random forest”for classification and regression[J].Chinese Journal of Applied Entomology,2013,50(4):1190-1197.)
• [13]钱忠东,李万.泄水锥形式对混流式水轮机压力脉动的影响分析[J].水力发电学报.2012,31(5):278-285.(QIAN Zhongdong,LI Wan.Analysis of pressure oscillation characteristics in Francis hydraulic turbine with different runner cones[J].Journal of Hydroelectric Engineering,2012,31(5):278-285.
• [14]BREIMAN,CUTLER A.Package random Forest[R/OL](2013-08-12)[2014-10-26].http://cran.r_project.org/web/packages/random Forest/randomForest.pdf.)