滑坡位移的多模态支持向量机模型预测

将支持向量机（support vector machine,SVM）方法与信号分析中的经验模态分解（empirical mode decomposition, EMD）方法相匹配,提出了一种通过多模态支持向量机函数回归分析建模预测滑坡位移的理论方法。以边坡位移历史观测数据为基础,应用EMD方法获得滑坡形成过程中位移演化的几个特征时间模态,构成了多模态信息统计学习样本,确定了边坡位移演化的自适应多尺度变化信息。对应于每个经验模态的位移变化信息,引入了多模态SVM建模方法,然后合成不同经验模态下边坡位移的计算结果,得到滑坡位移的预测值。以卧龙寺新滑坡和新滩滑坡的监测数据为基础的理论预测结果表明,与采用遗传算法的神经网络方法的预测结果相比,支持向量机经验模态方法具有更强的预测能力,理论预测结果与实际监测值具有很好的一致性     

加权支持向量回归机及其在水质预测中的应用

支持向量机是一种基于结构风险最小化原理的学习技术，也是一种新的具有很好泛化性能的回归方法。本文对用于回归估计的标准支持向量机加以改进，提出了一种新的用于回归估计的支持向量机学习算法，针对各样本重要性的差异，给各个样本的惩罚系数和误差要求赋予不同权重，并利用加权支持向量回归机的理论及其算法构建水质预测模型。实验结果表明，该方法对水质具有较好的预测效果。     

基于时间序列与人工蜂群支持向量机的滑坡位移预测研究

总结以往滑坡预测方法存在的诸多不足,针对滑坡监测位移-时间曲线特点,本文提出了一种基于时间序列的人工蜂群算法（ABC）与支持向量回归机（SVR）相结合的滑坡位移预测方法。以三峡库区白水河滑坡为例,通过对滑坡位移、降雨、库水位等因素的分析,研究影响滑坡位移变化的因素。用时间序列加法模型和移动平均法将滑坡位移分解为趋势项和周期项。以多项式最小二乘法拟合滑坡位移趋势项,用人工蜂群支持向量机模型对滑坡位移周期项进行训练和预测。通过灰色系统关联分析法计算多项因子与滑坡位移周期项之间的关联性。最终的滑坡总位移预测值为周期项预测值与趋势项预测值之和。与BP神经网络、PSO-SVR模型方法相比,该方法在滑坡位移预测中有更高的精度,在防灾减灾工作中有较好的推广应用前景。     

逻辑回归与支持向量机模型在滑坡敏感性评价中的应用

白龙江流域是我国滑坡泥石流灾害四大高发区之一,进行该区域滑坡敏感性评价,能够为决策者在灾害管理和设施建设规划方面提供帮助,对区域防灾减灾具有重要指导意义。本研究采用边坡单元为基本研究单元,在野外调查及前人研究基础上,选择控制该区域滑坡发育的19个要素作为影响因子;经过主成分分析和独立性检验得到该区域对滑坡形成贡献最大的6个因子:高程、坡度、坡向、岩性、断裂距离和人口密度;分别使用二元逻辑回归模型(LR)和支持向量机模型(SVM)对该区域进行滑坡敏感性评价;最后,采用ROC曲线对模型精度进行验证。研究结果表明,两模型各能将38.76%、14.48%、9.40%、11.28%、26.07%和13.49%、21.61%、8.17%、26.70%、30.04%的边坡单元分别预测为极高危险区、高危险区、中度危险区、低危险区和极低危险区;精度验证结果表明两种模型均能有效地进行该区域滑坡敏感性评价,并且支持向量机模型具有更好的分类能力、预测精度和稳定性。     

逻辑回归与支持向量机模型在滑坡敏感性评价中的应用

白龙江流域是我国滑坡泥石流灾害四大高发区之一,进行该区域滑坡敏感性评价,能够为决策者在灾害管理和设施建设规划方面提供帮助,对区域防灾减灾具有重要指导意义。本研究采用边坡单元为基本研究单元,在野外调查及前人研究基础上,选择控制该区域滑坡发育的19个要素作为影响因子; 经过主成分分析和独立性检验得到该区域对滑坡形成贡献最大的6个因子：高程、坡度、坡向、岩性、断裂距离和人口密度; 分别使用二元逻辑回归模型（LR）和支持向量机模型（SVM）对该区域进行滑坡敏感性评价; 最后,采用ROC曲线对模型精度进行验证。研究结果表明,两模型各能将38.76%、14.48%、9.40%、11.28%、26.07%和13.49%、21.61%、8.17%、26.70%、30.04%的边坡单元分别预测为极高危险区、高危险区、中度危险区、低危险区和极低危险区; 精度验证结果表明两种模型均能有效地进行该区域滑坡敏感性评价,并且支持向量机模型具有更好的分类能力、预测精度和稳定性。     

非线性支持向量机在油气预测中的应用

油气预测的传统方法通常是基于经验风险最小化准则,但在有限样本情况下,预测效果并不理想.研究引入基于结构化风险最小化准则的非线性支持向量机方法,通过对推广误差界的最小化达到最大的泛化能力和全局最优,时于小样本数据,该方法具有可靠的预测能力.在对四川观音场构造面新统上部碳酸盐岩储层数据处理中.通过实例试算,结果表明该方法有效可靠,预测精度高,与已知结果吻合较好.     

滑坡灾害空间预测研究

介绍了滑坡灾害空间预测的常用理论和方法、研究特点和适用范围，指出了区域滑坡空间预测、单体斜坡稳定性预测和滑坡灾害风险研究的发展趋势；以三峡库区巴东县黄土坡区斜坡稳定性区划为例，用神经网络模型和信息量模型两种方法进行了斜坡稳定性预测，取得了满意的效果。     

公路软基沉降预测的支持向量机模型

提出了基于支持向量机（SVM）模型对公路软基沉降进行预测的一种新方法,工程实例预测结果表明,在同样的训练均方误差下,SVM模型预测能力要优于BP神经网络模型,同时该模型能够综合利用分级加载过程中的沉降观测数据作为训练样本集,比仅依靠预压期内部分实测沉降数据的双曲线法更能反映地基土的变形趋势。因此,将建立的SVM模型应用于公路软基沉降预测能够更准确地反映实际沉降过程     

基于信息量法和支持向量机的芦山县滑坡危险性评价

地质滑坡对人员和经济形成了巨大威胁,因此,区域地质滑坡危险性评价变得极其重要.为了将滑坡带来的损失降到最低,这里以芦山县作为研究区,分析研究区内地质环境的滑坡成因等相关特征,选择12种滑坡因子作为其滑坡危险性评价指标,采用AHP层次分析法确定滑坡因子权重,使用基于径向基核函数的支持向量机模型和信息量模型,对芦山区域进行...     

Landslide susceptibility assessment using SVM machine learning algorithm

This paper introduces the current machine learning approach to solving spatial modeling problems in the domain of landslide susceptibility assessment. The latter is introduced as a classification problem, having multiple (geological, morphological, environmental etc.) attributes and one referent landslide inventory map from which to devise the classification rules. Three different machine learning algorithms were compared: Support Vector Machines, Decision Trees and Logistic Regression. A specific area of the Fruška Gora Mountain (Serbia) was selected to perform the entire modeling procedure, from attribute and referent data preparation/processing, through the classifiers' implementation to the evaluation, carried out in terms of the model's performance and agreement with the referent data. The experiments showed that Support Vector Machines outperformed the other proposed methods, and hence this algorithm was selected as the model of choice to be compared with a common knowledge-driven method – the Analytical Hierarchy Process – to create a landslide susceptibility map of the relevant area. The SVM classifier outperformed the AHP approach in all evaluation metrics (κ index, area under ROC curve and false positive rate in stable ground class).     

Landslide susceptibility mapping using GIS-based multi-criteria decision analysis,support vector machines,and logistic regression

Identification of landslides and production of landslide susceptibility maps are crucial steps that can help planners, local administrations, and decision makers in disaster planning. Accuracy of the landslide susceptibility maps is important for reducing the losses of life and property. Models used for landslide susceptibility mapping require a combination of various factors describing features of the terrain and meteorological conditions. Many algorithms have been developed and applied in the literature to increase the accuracy of landslide susceptibility maps. In recent years, geographic information system-based multi-criteria decision analyses (MCDA) and support vector regression (SVR) have been successfully applied in the production of landslide susceptibility maps. In this study, the MCDA and SVR methods were employed to assess the shallow landslide susceptibility of Trabzon province (NE Turkey) using lithology, slope, land cover, aspect, topographic wetness index, drainage density, slope length, elevation, and distance to road as input data. Performances of the methods were compared with that of widely used logistic regression model using ROC and success rate curves. Results showed that the MCDA and SVR outperformed the conventional logistic regression method in the mapping of shallow landslides. Therefore, multi-criteria decision method and support vector regression were employed to determine potential landslide zones in the study area.     

Landslide susceptibility mapping using GIS and digital photogrammetric techniques: a case study from Ardesen (NE-Turkey)

Ardesen is a settlement area which has been significantly damaged by frequent landslides which are caused by severe rainfalls and result in many casualties. In this study a landslide susceptibility map of Ardesen was prepared using the Analytical Hierarchy Process (AHP) with the help of Geographical Information Systems (GIS) and Digital Photogrametry Techniques (DPT). A landslide inventory, lithology–weathering, slope, aspect, land cover, shear strength, distance to the river, stream density and distance to the road thematics data layers were used to create the map. These layer maps are produced using field, laboratory and office studies, and by the use of GIS and DPT. The landslide inventory map is also required to determine the relationship between these maps and landslides using DPT. In the study field in the Hemsindere Formation there are units that have different weathering classes, and this significantly affects the shear strength of the soil. In this study, shear strength values are calculated in great detail with field and laboratory studies and an additional layer is evaluated with the help of the stability studies used to produce the landslide susceptibility map. Finally, an overlay analysis is carried out by evaluating the layers obtained according to their weight, and the landslide susceptibility map is produced. The study area was classified into five classes of relative landslide susceptibility, namely, very low, low, moderate, high, and very high. Based on this analysis, the area and percentage distribution of landslide susceptibility degrees were calculated and it was found that 28% of the region is under the threat of landslides. Furthermore, the landslide susceptibility map and the landslide inventory map were compared to determine whether the models produced are compatible with the real situation resulting in compatibility rate of 84%. The total numbers of dwellings in the study area were determined one by one using aerial photos and it was found that 30% of the houses, with a total occupancy of approximately 2,300 people, have a high or very high risk of being affected by landslides.     

Landslide susceptibility mapping of vicinity of Yaka Landslide (Gelendost,Turkey) using conditional probability approach in GIS

On 19 February 2007, a landslide occurred on the Alaard?ç Slope, located 1.6 km south of the town of Yaka (Gelendost, Turkey.) Subsequently, the displaced materials transformed into a mud flow in E?lence Creek and continued 750 m downstream towards the town of Yaka. The mass poised for motion in the Yaka Landslide source area and its vicinity, which would be triggered to a kinetic state by trigger factors such as heavy or sustained rainfall and/or snowmelt, poise a danger in the form of loss of life and property to Yaka with its population of 3,000. This study was undertaken to construct a susceptibility mapping of the vicinity of the Yaka Landslide’s source area and to relate it to movement of the landslide mass with the goal of prevention or mitigation of loss of life and property. The landslide susceptibility map was formulated by designating the relationship of the effecting factors that cause landslides such as lithology, gradient, slope aspect, elevation, topographical moisture index, and stream power index to the landslide map, as determined by analysis of the terrain, through the implementation of the conditional probability method. It was determined that the surface area of the Goksogut formation, which has attained lithological characteristics of clayey limestone with a broken and separated base and where area landslides occur, possesses an elevation of 1,100–1,300 m, a slope gradient of 15 °–35 ° and a slope aspect between 0 °–67.5 ° and 157 °–247 °. Loss of life and property may be avoided by the construction of structures to check the debris mass in E?lence Creek, the cleaning of the canal which passes through Yaka, the broadening of the canal’s base area, elevating the protective edges along the canal and the establishment of a protective zone at least 10-m wide on each side of the canal to deter against damage from probable landslide occurrence and mud flow.     

Integrated Landslide Susceptibility Analysis and Hazard Assessment in the Principality of Andorra

In the paper we present the procedure for hazard assessment that has been used to prepare the landslide hazard map of the Principality of Andorra at 1:5,000 scale. The main phases of the hazard assessment are discussed. Susceptibility analysis has involved the location of the potential slope failures, and the estimation of both landslide volume and runout distance. In the susceptible areas, landslide magnitude and frequency has been determined in order to produce the Hazard Zoning Map. Data required for hazard assessment have been introduced into a GIS or derived directly from available Digital Terrain Models. Data handling and treatment with the GIS has allowed the performance of the landslide hazard assessment and mapping in a fast and reproducible way.     

Prediction of ultimate axial load-carrying capacity of piles using a support vector machine based on CPT data

The support vector machine (SVM) is a relatively new artificial intelligence technique which is increasingly being applied to geotechnical problems and is yielding encouraging results. In this paper SVM models are developed for predicting the ultimate axial load-carrying capacity of piles based on cone penetration test (CPT) data. A data set of 108 samples is used to develop the SVM models. These data were obtained from the literature containing pile load tests and each sample contains information regarding pile geometry, full-scale static pile load tests and CPT results. Moreover, a sensitivity analysis is carried out to examine the relative significance of each input variable with respect to ultimate strength prediction. Finally, a statistical analysis is conducted to make comparisons between predictions obtained from the SVM models and three traditional CPT-based methods for determining pile capacity. The comparison confirms that the SVM models developed in this paper outperform the traditional methods.     

边坡可靠度分析的支持向量机法

提出了基于支持向量机（SVM）的边坡可靠度分析新算法。该方法采用均匀设计确定样本点,通过一定数量的确定性计算来训练SVM,拟合边坡的功能函数;采用一阶可靠度方法（FORM）和迭代算法优化SVM模型,获得可靠度指标和验算点信息;在SVM模型基础上进一步通过二阶可靠度方法（SORM）和蒙特卡罗模拟（MCS）计算边坡的失稳概率。以两个典型边坡为例,通过与其他方法比较,证明了该方法的准确性和高效性。结果表明：提出的在标准正态空间（U空间）中取样并构建SVM,在原始空间（X空间）中计算功能函数的算法,有效地解决了具有相关非正态分布变量的可靠度分析问题,并且可很容易扩展到SORM的计算。算例结果证明,该方法的精度高于FORM;而效率优于MCS。分析过程中,边坡安全系数计算和可靠度分析相互独立。因此,该方法既适用于具有显式功能函数的简单问题,也适用于需要软件计算安全系数的实际边坡问题。     

基于SOM-I-SVM耦合模型的滑坡易发性评价

在使用机器学习模型对滑坡进行易发性评价时,通常会在滑坡影响范围之外随机选取非滑坡样本点,具有一定的误差。为了提高滑坡易发性评价的精度,将自组织映射（self-organizing map,SOM）神经网络、信息量模型（information,I）以及支持向量机模型（support vector machine,SVM）进行耦合,提出一种基于SOM-I-SVM模型的滑坡易发性评价方法,并将SOM神经网络与K均值聚类算法进行对比,验证模型的可靠性。以十堰市茅箭区为例,首先通过对环境因子的相关性及重要性分析,筛选出距水系距离、坡度、降雨量、距构造距离、相对高差、距道路距离、地层岩性等7个因子,建立滑坡易发性评价指标体系,在此基础上计算出各因子的分级信息量值,并作为模型的输入变量进行滑坡易发性评价。分别采用SOM神经网络和K均值聚类算法选取非滑坡样本,然后将样本数据集代入I-SVM模型预测滑坡易发性。将SVM、I-SVM、KMeans-I-SVM、SOM-I-SVM等4种模型预测精度进行对比,其ROC曲线下面积（AUC）分别为0.82,0.88,0.90,0.91,说明SOM-I-SVM模型能有效提高滑坡易发性预测准确率。     

基于XGBoost模型的三峡库区燕山乡滑坡易发性评价与区划

滑坡易发性评价是精细化滑坡灾害风险评价的基础。为了提升滑坡易发性评价模型的精度和稳健性,以三峡库区万州区燕山乡为例,选取工程地质岩组、堆积层厚度等九个影响因子构建滑坡易发性评价指标体系,应用信息量模型定量分析滑坡发育与指标之间的关系。在此基础上,随机选取70%/30%的滑坡样本作为训练/验证数据集,应用极致梯度提升模型(extreme gradient boosting, XGBoost)开展易发性评价。随后从模型预测精度和模型稳定性两方面将其与决策树模型(decision tree, DT)和梯度提升树模型(gradient boosting decision tree, GBDT)进行对比。结果表明：研究区堆积层滑坡主要受长江水系、堆积层厚度和工程地质岩组影响。XGBoost模型具有最高的准确率(94.3%)和预测精度(97.3%)。在模型稳定性验证中,平均预测精度最高(97.3%),优于DT(91.3%)和GBDT(95.7%),模型标准差和变异系数均为0.01,低于其余两种模型。XGBoost在区域滑坡易发性评价与制图中得到了可靠的结果,为滑坡灾害空间预测提供了新的技术支撑。