处理非线性分类和回归问题的一种新方法（Ⅰ）——支持向量机方法简介   总被引：51，自引：3，他引：51       下载免费PDF全文

陈永义俞小鼎  高学浩冯汉中 《应用气象学报》2004,15(3):345-354

简要介绍了近年来倍受瞩目的一种处理高度非线性分类、回归等问题的计算机学习的新方法——支持向量机(SVM)方法；分析了这一方法的特点及其在数值预报产品释用及气象研究业务中的应用前景。SVM是一种有坚实理论基础的新颖的小样本学习方法。它基本上不涉及概率测度及大数定律等，因此不同于现有的统计方法。从本质上看，它避开了从归纳到演绎的传统过程，实现了高效的从训练样本到预报样本的“转导推理”(transductive inference)，大大简化了通常的分类和回归等问题。SVM的最终决策函数只由少数的支持向量所确定，计算的复杂性取决于支持向量的数目，而不是样本空间的维数，这在某种意义上避免了“维数灾”。  相似文献   

A note on the statistics of the largest geomagnetic storms per solar cycle     

A.R. Solow  A.R. Beet 《Journal of Atmospheric and Solar》2004,66(18):1731

This note points out a problem with the way in which extreme value distributions have been fit to the intensities of the largest geomagnetic storms per solar cycle. An alternative method is described. This method is applied to observations of the three largest geomagnetic storms in solar cycles 11–22.  相似文献   

江苏省月平均最高、最低气温周期振动的谱特征   总被引：3，自引：3，他引：3       下载免费PDF全文

何卷雄  姜爱军等 《气象科学》2002,22(2):159-166

本文综合运用奇异交叉谱分析等多种气候诊断方法研究江苏月平均最高、最低气温振动的准周期变化的时空分布特征。结果发现：江苏逐月最高、最低气温分布型态分别为经向型、纬向型；两者存在2至3年周期，显著耦合周期为3．5年及16．5年。本文同时讨论了奇异交叉谱分析、小波分析、最大熵谱分析、经典交叉谱分析的优缺点。  相似文献   

A proposed mechanism for storm beach sedimentation     

Julian D. Orford 《地球表面变化过程与地形》1977,2(4):381-400

One of the major problems of shingle beach dynamics is the method by which coarse gravel is transferred from beach face to storm beach, the latter often lying several metres above high spring tidal levels. The mechanism which is usually proposed, cites the action of plunging breakers as being central to this problem of sediment transfer. However, the nature of net residual fluid force of plunging breakers is deemed unsatisfactory for any substantial upbeach (onshore) sediment transport during storm conditions on gravel beaches. A mechanism is proposed by which high still water levels due to high astronomical tides, onshore storm force winds and allied wave surge, promote shoaling characteristics and beach profile changes which are conducive to spilling breaker development at tidal extremities. It is the net onshore fluid force vector of the spilling breaker overtopping the beach crest which may be the cause of extreme sedimentation events on the storm beach. An example of such sediment transfer is given for a known storm beach sedimentation event at Llanrhystyd gravel beach, West Wales, during February 1974. Process variables were monitored on several days allowing the use of an inshore breaker steepness criterion, to indicate positions in the tidal regime where plunging breakers give way to spilling forms. This example serves to suggest that more attention should be given to the nature and characteristics of shoaling waves, especially in respect of breaker type, when examining problems of shingle beach dynamics and sedimentation.  相似文献   

Extreme rainfall events in Catalonia. The case of 12 November 1988     

M. Carmen Llasat  Roberto Rodriguez 《Natural Hazards》1992,5(2):133-151

A study of the synoptic situation which produced the catastrophic floods of November 1988 in Catalonia (in the northeast of the Iberian Peninsula) is presented. Analyses of the vertical structure, potential instability, precipitable water, and instability index are made through the radiosounding data from Palma, Majorca. It is found that the 1988 situation is included in type I intense convective events in Catalonia (classification obtained from all the events since 1950, (Llasat, 1989)). It was characterized by:

(a)	-pattern in the middle and high troposphere, the ridge axis east of Catalonia.
(b)	High pressure over Europe.
(c)	South-easterly winds in the lower troposphere with warm and moist humid air advection and south-westerlies aloft over Catalonia.
(d)	Strong instability (convective and latent).
(e)	Penetration of Atlantic air.

  相似文献   

胡杨根系分布特征与根系吸水模型建立   总被引：16，自引：1，他引：15  

冯起  司建华  李建林  席海洋 《地球科学进展》2008,23(7):765-772

依据2006年和2007年6~7月在极端干旱区额济纳的实测资料,利用分形和概率统计的方法对胡杨根系分布与根区土壤水分的关系进行了分析研究。引入土壤含水率期望(cm·g/g)的概念,建立了含水率期望值与根系分维值之间的函数关系,并求出了适宜胡杨生长的土壤含水率期望值的范围。同时建立了初版的根系吸水模型,并进行验证,得到较好的结果。  相似文献   

Governing fisheries as complex adaptive systems     

Robin Mahon  Patrick McConney  Rathindra N. Roy 《Marine Policy》2008

Fisheries are complex human-in-nature systems. The conventional approach to fisheries systems has been to treat them as predictable and controllable. As complex systems they are neither of the two and have to be approached differently. Complex systems often exhibit the capacity to self-organize or adapt, even without outside influence. If this is true of fisheries, it should lead to a radically different approach to management of fisheries systems that places much emphasis on enabling self-organization, learning and adaptation. Conceptual and practical frameworks for enabling activities are needed.  相似文献   

基于主成分-正则化极限学习机的超高密度电法非线性反演   总被引：2，自引：0，他引：2       下载免费PDF全文

江沸菠  戴前伟  董莉 《地球物理学报》2015,58(9):3356-3369

超高密度电法是一种新的地球物理探测技术,它通过多通道数据采集和多装置数据联合反演,极大地提高了电法勘探的成像精度.本文提出一种主成分-正则化极限学习机(PC-RELM)非线性反演方法,该方法针对超高密度电法所获取的高维勘探数据进行反演建模,通过随机设定隐层参数来简化模型的学习过程,通过主成分分析方法来进行高维数据降维,最后引入正则化因子提高反演模型的泛化能力.论文给出了超高密度电法的原理、样本构造方法和非线性反演流程,使用交叉验证方法获得了优化的隐节点数目和正则化参数,构造了优化的反演模型.通过两个经典的超高密度模型的反演结果表明,该方法能够较好地解决超高密度电法反演的高维数据非线性建模问题,能够弥补单一装置数据反演的不足,同时相较其他的非线性反演方法(ELM,BPNN和GRNN)具有更加准确的反演结果.  相似文献   

基于土壤反射光谱特性的广东省稻田土壤快速分类   总被引：1，自引：0，他引：1  

李丹  彭智平  韩留生  王重洋  刘尉  黄思宇  陈水森 《热带地理》2015,35(1):29-34

选择广东省215个村镇稻田的土壤样本,首先利用ASD Field Spec3测量土壤样本在350~2 500 nm的光谱,并采用S-G一阶导数平滑滤波降低样本测量中光照差异的影响,然后将遗传算法（Genetic Algorithm,GA）和支持向量机分类（Support Vector Machine,SVM）分别用于提取分类光谱特征和建立分类模型,分别在土纲、亚纲、土类3个层次进行土壤分类。结果表明：1）在不同的分类层次下,与铁氧化物密切相关的650~710以及900 nm附近光谱,与羟基矿物吸收有关的2 207~2 237和2 377~2 397 nm区间均被作为分类特征变量。2）随着土壤类型的细分,分类所需变量增多。在土类级,对有机质敏感的2 080 nm附近的光谱也被引入分类定标模型中,土纲和亚纲下分类精度＞67%,土类级分类精度为58.67%。利用遗传算法提取光谱特征,进行基于支持向量机的土壤分类具有一定优势。  相似文献   

湖南省极端降水过程客观监测指标     

廖玉芳  张剑明  刘兵  李英  唐爱军  段丽洁 《气象与环境学报》2014,30(5):146-152

利用1981-2010年湖南省97个地面气象观测站降水观测资料,基于二阶求导和尖顶突变模型等方法,分别对短历时降水序列和过程降水序列进行时间突变分析。结果表明：不同时间的最大小时累积降水量在时间上的突变点分别为12.0 h、6.1 h;过程降水持续时间突变点分别为10.0 d、4.4 d,不同持续时间最大降水量占过程总降水量的百分比及最大降水量在时间序列上的突变点分别为10.0 d、4.6 d和10.0 d、4.9d。建立的湖南省极端降水过程客观监测指标为：连续1-6 h、12 h最大降水量,连续1-5 d、10 d最大降水量和过程累积降水量及降水持续时间。指标的适用性通过与农作物因洪涝灾害的受灾率、成灾率及单位面积粮食减产量的相关分析验证。  相似文献