共查询到20条相似文献，搜索用时 31 毫秒
1.
A test of the ability of a probabilistic neural network to classify deposits into types on the basis of deposit tonnage and average Cu, Mo, Ag, Au, Zn, and Pb grades is conducted. The purpose is to examine whether this type of system might serve as a basis for integrating geoscience information available in large mineral databases to classify sites by deposit type. Benefits of proper classification of many sites in large regions are relatively rapid identification of terranes permissive for deposit types and recognition of specific sites perhaps worthy of exploring further.Total tonnages and average grades of 1,137 well-explored deposits identified in published grade and tonnage models representing 13 deposit types were used to train and test the network. Tonnages were transformed by logarithms and grades by square roots to reduce effects of skewness. All values were scaled by subtracting the variable's mean and dividing by its standard deviation. Half of the deposits were selected randomly to be used in training the probabilistic neural network and the other half were used for independent testing. Tests were performed with a probabilistic neural network employing a Gaussian kernel and separate sigma weights for each class (type) and each variable (grade or tonnage).Deposit types were selected to challenge the neural network. For many types, tonnages or average grades are significantly different from other types, but individual deposits may plot in the grade and tonnage space of more than one type. Porphyry Cu, porphyry Cu-Au, and porphyry Cu-Mo types have similar tonnages and relatively small differences in grades. Redbed Cu deposits typically have tonnages that could be confused with porphyry Cu deposits, also contain Cu and, in some situations, Ag. Cyprus and kuroko massive sulfide types have about the same tonnages. Cu, Zn, Ag, and Au grades. Polymetallic vein, sedimentary exhalative Zn-Pb, and Zn-Pb skarn types contain many of the same metals. Sediment-hosted Au, Comstock Au-Ag, and low-sulfide Au-quartz vein types are principally Au deposits with differing amounts of Ag.Given the intent to test the neural network under the most difficult conditions, an overall 75% agreement between the experts and the neural network is considered excellent. Among the largestclassification errors are skarn Zn-Pb and Cyprus massive sulfide deposits classed by the neuralnetwork as kuroko massive sulfides—24 and 63% error respectively. Other large errors are the classification of 92% of porphyry Cu-Mo as porphyry Cu deposits. Most of the larger classification errors involve 25 or fewer training deposits, suggesting that some errors might be the result of small sample size. About 91% of the gold deposit types were classed properly and 98% of porphyry Cu deposits were classes as some type of porphyry Cu deposit. An experienced economic geologist would not make many of the classification errors that were made by the neural network because the geologic settings of deposits would be used to reduce errors. In a separate test, the probabilistic neural network correctly classed 93% of 336 deposits in eight deposit types when trained with presence or absence of 58 minerals and six generalized rock types. The overall success rate of the probabilistic neural network when trained on tonnage and average grades would probably be more than 90% with additional information on the presence of a few rock types. 相似文献
2.
In order to determine whether it is desirable to quantify mineral-deposit models further, a test of the ability of a probabilistic
neural network to classify deposits into types based on mineralogy was conducted. Presence or absence of ore and alteration
mineralogy in well-typed deposits were used to train the network. To reduce the number of minerals considered, the analyzed
data were restricted to minerals present in at least 20% of at least one deposit type. An advantage of this restriction is
that single or rare occurrences of minerals did not dominate the results. Probabilistic neural networks can provide mathematically
sound confidence measures based on Bayes theorem and are relatively insensitive to outliers. Founded on Parzen density estimation,
they require no assumptions about distributions of random variables used for classification, even handling multimodal distributions.
They train quickly and work as well as, or better than, multiple-layer feedforward networks. Tests were performed with a probabilistic
neural network employing a Gaussian kernel and separate sigma weights for each class and each variable. The training set was
reduced to the presence or absence of 58 reported minerals in eight deposit types. The training set included: 49 Cyprus massive
sulfide deposits; 200 kuroko massive sulfide deposits; 59 Comstock epithermal vein gold districts; 17 quartzalunite epithermal
gold deposits; 25 Creede epithermal gold deposits; 28 sedimentary-exhalative zinc-lead deposits; 28 Sado epithermal vein gold
deposits; and 100 porphyry copper deposits. The most common training problem was the error of classifying about 27% of Cyprus-type
deposits in the training set as kuroko. In independent tests with deposits not used in the training set, 88% of 224 kuroko
massive sulfide deposits were classed correctly, 92% of 25 porphyry copper deposits, 78% of 9 Comstock epithermal gold-silver
districts, and 83% of six quartzalunite epithermal gold deposits were classed correctly. Across all deposit types, 88% of
deposits in the validation dataset were correctly classed. Misclassifications were most common if a deposit was characterized
by only a few minerals, e.g., pyrite, chalcopyrite,and sphalerite. The success rate jumped to 98% correctly classed deposits
when just two rock types were added. Such a high success rate of the probabilistic neural network suggests that not only should
this preliminary test be expanded to include other deposit types, but that other deposit features should be added 相似文献
3.
The need to integrate large quantities of digital geoscience information to classify locations as mineral deposits or nondeposits has been met by the weights-of-evidence method in many situations. Widespread selection of this method may be more the result of its ease of use and interpretation rather than comparisons with alternative methods. A comparison of the weights-of-evidence method to probabilistic neural networks is performed here with data from Chisel Lake-Andeson Lake, Manitoba, Canada. Each method is designed to estimate the probability of belonging to learned classes where the estimated probabilities are used to classify the unknowns. Using these data, significantly lower classification error rates were observed for the neural network, not only when test and training data were the same (0.02 versus 23%), but also when validation data, not used in any training, were used to test the efficiency of classification (0.7 versus 17%). Despite these data containing too few deposits, these tests of this set of data demonstrate the neural network's ability at making unbiased probability estimates and lower error rates when measured by number of polygons or by the area of land misclassified. For both methods, independent validation tests are required to ensure that estimates are representative of real-world results. Results from the weights-of-evidence method demonstrate a strong bias where most errors are barren areas misclassified as deposits. The weights-of-evidence method is based on Bayes rule, which requires independent variables in order to make unbiased estimates. The chi-square test for independence indicates no significant correlations among the variables in the Chisel Lake–Andeson Lake data. However, the expected number of deposits test clearly demonstrates that these data violate the independence assumption. Other, independent simulations with three variables show that using variables with correlations of 1.0 can double the expected number of deposits as can correlations of –1.0. Studies done in the 1970s on methods that use Bayes rule show that moderate correlations among attributes seriously affect estimates and even small correlations lead to increases in misclassifications. Adverse effects have been observed with small to moderate correlations when only six to eight variables were used. Consistent evidence of upward biased probability estimates from multivariate methods founded on Bayes rule must be of considerable concern to institutions and governmental agencies where unbiased estimates are required. In addition to increasing the misclassification rate, biased probability estimates make classification into deposit and nondeposit classes an arbitrary subjective decision. The probabilistic neural network has no problem dealing with correlated variables—its performance depends strongly on having a thoroughly representative training set. Probabilistic neural networks or logistic regression should receive serious consideration where unbiased estimates are required. The weights-of-evidence method would serve to estimate thresholds between anomalies and background and for exploratory data analysis. 相似文献
4.
基于主成分分析的BP神经网络在城市建成区面积预测中的应用——以北京市为例 总被引：1，自引：0，他引：1
城市建成区规模受社会、经济、城市环境等诸多因素影响, 传统统计方法难以准确预测城 市建成区的面积。人工神经网络具有良好的非线性映射逼近性能, 在各类预测研究中得到了广泛 的应用, 尤其是BP 神经网络。主成分分析可以在有效保留数据信息前提下对数据进行降维, 它 与BP 神经网络的结合主要在数据输入端, 通过减少输入层神经元个数, 增强网络性能, 提高预 测精度。本文以北京市为例, 综合运用主成分分析和BP 神经网络方法建立预测模型, 以1986~ 2003 年数据为学习样本, 以2004 年数据为检验样本, 对2005 年北京市城市建成区面积进行模 拟预测。预测结果表明, 基于主成分分析的BP 神经网络预测结果与实际值的相对误差为2.8%, 比传统BP 神经网络预测精度提高1.8 个百分点, 网络训练收敛速度也更快, 其预测精度和效率 都有不同程度的改善。 相似文献
5.
基于空间滤波方法的中国省际人口迁移驱动因素 总被引：9，自引：5，他引：4
人口迁移数据中往往存在较强的网络自相关性,以往基于最小二乘估计的重力模型与迁移数据的拟合度较低,而改进后的泊松重力模型仍存在过度离散的缺陷,以上问题均导致既有人口迁移模型中的估计偏差。本文构建了特征向量空间滤波（ESF）负二项重力模型,基于2015年全国1%人口抽样调查数据,研究2010-2015年中国省际人口迁移的驱动因素。结果表明：① 省际人口迁移流间存在显著的空间溢出效应,ESF能有效地提取数据中的网络自相关性以降低模型的估计偏差,排序在前1.4%的特征向量即可提取较强的网络自相关信息。② 省际人口迁移流之间存在明显的过度离散现象,考虑到数据离散的负二项重力模型更适用于人口迁移驱动因素的估计。③ 网络自相关性会导致模型对距离相关变量估计的上偏与大部分非距离变量估计的下偏,修正后的模型揭示出以下驱动因素：区域人口特征、社会网络、经济发展、教育水平等因素是引发省际人口迁移的重要原因,而居住环境与公路网络等因素也逐渐成为影响人口迁移重要的“拉力”因素。④ 与既有研究相比,社会网络因素(迁移存量、流动链指数)对人口迁移的影响日益增强,而空间距离对人口迁移的影响进一步呈现弱化趋势。 相似文献
6.
The objective of this project was to characterize the freeze-thaw properties of recycled concrete (RCA) and asphalt (RAP) as unbound base and to assess how they behaved in the field for nearly 8 years. This paper includes an examination of existing information, laboratory studies of freeze-thaw behavior, and evaluation of data from MnROAD field-test sections in a seasonally cold region, i.e., in Minnesota, USA. Test sections were constructed using recycled materials in the granular base layers at the MnROAD test facility. One test section included 100% RAP, another 100% RCA, a third one a 50/50 blend of RCA/natural aggregate, and a fourth one only natural aggregate (Class 5) as a control. The stiffness (i.e., elastic modulus) was monitored during construction and throughout the pavement life by the Minnesota Department of Transportation, along with the variation of temperatures and moisture regimes in the pavement to determine their effects on pavement performance. The resilient modulus of each material was determined by bench-scale testing in accordance with NCHRP 1-28a, as well as by field-scale tests incorporating a falling-weight deflectometer. Specimens were subjected to as many as 20 cycles of freeze-thaw in the laboratory, and the change in their resilient modulus was measured. In the field-test sections constructed with the same materials as the base course, temperature, moisture, and field modulus (from falling-weight deflectometer tests) were monitored seasonally for nearly 8 years. From the temperatures in the base course layer, the number of freeze-thaw cycles experienced in the field was determined for each test section. Inferences were made relative to modulus change versus freeze-thaw cycles. Conclusions were drawn for long-term field performances of the recycled base (RAB) in comparison to natural aggregate. 相似文献
7.
The relationship between seismic velocity and density in the continental crust — a useful constraint? 总被引：1，自引：0，他引：1
P. J. Barton 《Geophysical Journal International》1986,87(1):195-208
Summary . Plots of seismic velocity and density of rock samples show that a range of densities is possible for rocks of each seismic velocity and vice versa. although a single linear relationship is often assumed in crustal gravity calculations. Because of the scatter, whenever rocks of known seismic velocity are converted to density using this relationship, a reduction is made to the resolving power of the resulting gravity calculation. If these rocks reach thicknesses of more than a few kilometres, then the uncertainties become significant when compared with the size of commonly observed gravity anomalies. Examples are considered from the North Sea, Mississippi and Carolina Trough. It is concluded that the use of a seismic velocity measurement as the only indication of rock density does not provide a useful constraint when attempting to reproduce observed gravity variations. An appropriate model for isostatic compensation is probably the most important factor for successful predictions of crustal structure on the basis of gravity data. 相似文献
8.
To investigate the static and dynamic resilient modulus of fine soil, and adapting to the new design code and maintenance system of highway subgrade in China, a series of static and dynamic tests were carried out according to the standard laboratory test methods (JTG E40-2007 and JTG D30-2015, respectively). The effects of initial water content, compactness and freeze-thaw cycles on the static and dynamic resilient moduli of fine soil were investigated and analyzed. Experimental test results show that with increasing water content, dry density and freeze-thaw cycles, the static moduli reduces about 10.2%~40.0%, 14.4%~45.5%, and 24.0%~50.3%, and dynamic moduli reduces about 10.9%~90.8%, 2.5%~38.4%, and 0.0%~46.0%, respectively. Then, the empirical mathematical relationship between static and dynamic resilient moduli was established under different water content, dry density and freeze-thaw cycles. The investigation results can be used to determine the dynamic modulus of fine soil by widely used static modulus, which could meet the requirement of adopting dynamic modulus index in new specification. 相似文献
9.
Freeze-thaw damage is the most common disease of semi-rigid bases in cold regions, which may greatly affect the durability of roadways. In this study, the compressive strength and frost resistance of four different types of semi-rigid bases (lime-fly ash-stabilized sand, cement-stabilized sand, lime-fly ash-stabilized gravel, and cement-stabilized gravel) are assessed by varying the materials content. Based on freeze-thaw and compressive strength tests, this paper presents the performance of the different materials, each having different physical properties, and the optimal amounts of materials contents are proposed. 相似文献
10.
针对陇中黄土丘陵沟壑区土壤水蚀过程复杂且难以有效预测的问题,以定西市安家沟水土保持试验站2005—2016年1~12月人工草地径流场试验数据为主要来源,将流域月降雨量、月侵蚀性降雨量、月径流量、月降雨强度、径流场面积、径流场坡度、土壤砂粒含量、土壤粘粒含量8个因子作为输入因子,月土壤水蚀量作为输出,运用偏最小二乘法(Partial Least-Squares
Regression,PLSR)和长短期记忆(Long Short-Term Memory,LSTM)循环神经网络建立人工草地土壤水蚀预测模型,并利用BP(Back Propagation)、RNN(Recurrent Neural
Network)、LSTM常见神经网络模型,对模型的有效性进行评估。结果表明：PLSR将模型8个输入因子减少为4个,从而有效解决LSTM神经网络模型对样本数量要求过高的问题; PLSR和LSTM神经网络模型的结合可以有效提高模型对人工草地土壤水蚀过程的预测精度和收敛速度,预测结果的平均相对误差小于4%,相关系数高于其他3种神经网络模型,而迭代次数、均方根误差和平均绝对误差均低于其他3种模型;研究发现坡度对人工草地土壤水蚀过程影响较为明显,降雨量小于25 mm时,人工草地土壤水蚀量不会随坡度增加而明显增长,但当降雨量超过25 mm时,人工草地土壤水蚀量会随坡度明显增加。 PLSR LSTM神经网络土壤水蚀预测模型可以准确预测陇中黄土丘陵沟壑区人工草地土壤水蚀量,为该地区水土流失的准确预报提供新的思路和方法。 相似文献
11.
The fold test as an analytical tool 总被引：1，自引：0，他引：1
P. L. McFadden 《Geophysical Journal International》1998,135(2):329-338
Two fold tests for palaeomagnetism have recently been proposed that rely on the assumption that the total population of magnetic vectors is most highly concentrated with the rocks in the orientation they had at the time of acquisition of the magnetization. This leads to appealing, simple tests based on parameter estimation. However, it is shown that the underlying assumption is flawed and can lead to incorrect conclusions. McFadden & Jones (1981 ) previously developed an inference test based on the concept that the between-group dispersion of magnetization should be consistent with the within-group dispersion when the rocks are in the orientation they had at the time of magnetic acquisition. That test made unrealistic demands upon the sampling scheme for typical, realistic folding geometries and so it has been under-utilized. The McFadden & Jones test is extended by recognizing that it is sufficient to use groups with similar bedding corrections and that it is not necessary to insist on groups with common bedding corrections. These groups may easily be determined with a clustering algorithm. The point is that with the rocks in the orientation at which the magnetization was acquired, it should be immaterial how the groups are chosen. 相似文献
12.
This paper aims to determine the optimal fines content of coarse-grained soil required to simultaneously achieve weaker frost susceptibility and better bearing capacity. We studied the frost susceptibility and strength properties of coarse-grained soil by means of frost heaving tests and static triaxial tests, and the results are as follows:(1) the freezing temperature of coarse-grained soil decreased gradually and then leveled off with incremental increases in the percent content of fines; (2) the fines content proved to be an important factor influencing the frost heave susceptibility and strength properties of coarse-grained soil. With incremental increases in the percent content of fines, the frost heave ratio increased gradually and the cohesion function of fines effectively enhanced the shear strength of coarse-grained soil before freeze-thaw, but the frost susceptibility of fines weakened the shear strength of coarse-grained soil after freeze-thaw; (3) with increasing numbers of freeze-thaw cycles,the shear strength of coarse-grained soil decreased and then stabilized after the ninth freeze-thaw cycle, and therefore the mechanical indexes of the ninth freeze-thaw cycle are recommended for the engineering design values; and (4) considering frost susceptibility and strength properties as a whole, the optimal fines content of 5% is recommended for railway subgrade coarse-grained soil fillings in frozen regions. 相似文献
13.
基于小波变换和GRNN神经网络的黑河出山径流模型 总被引：14，自引：6，他引：8
对黑河山区流域月降水量和气温做Harr小波变换,并作为GRNN神经网络的输入,对黑河出山径流进行模拟和预测验证,效果较好。应用全球变化成果,在不同的气候情景下,对黑河出山径流进行预测。结果表明,黑河出山径流在未来一段时间内,径流量会有一定程度的增加,最终会减少。但模型对气温反应不敏感。去除气温重构的细节系数后,气温也成为一个敏感因素,但径流量却随气温的增加而增加。可推断,引进Haar小波变换的GRNN神经网络模型可应用于径流量对气温不敏感的流域。 相似文献
14.
Finiteelement simulations are increasingly providing a versatile environment for this topic. In this study, a two-dimensional finite element analysis is conducted to predict the deformation of highembankment in Bazhun heavy-haul railway, China. A recently developed nonlinear softening-type constitutive model is utilized to model the behavior of subgrade filling materials subjected to freeze-thaw cycles. For the convenience of practical application, the dynamic loading induced by a vehicle is treated as a quasi-static axle load. The deformation of this embankmentwith different moisture content under freeze-thaw cycles is compared. The results show that when subjected to the first freeze-thaw cycle, the embankmentexperienced significant deformation variations. Maximum deformation was usually achieved after the embankment with optimum moisture content experienced six freeze-thaw cycles, however, the embankment with moisture content of 8.0% and 9.5% deforms continuously even after experiencing almost ten freeze-thaw cycles. Overall, this study provides a simple nonlinear finite element approach for calculating the deformation of the embankmentinchanging climate conditions. 相似文献
15.
J. Urrutia Fucugauchi 《Geophysical Journal International》1980,61(2):243-251
Summary. In palaeomagnetic studies of volcanic rocks it is often considered that, if the direction of NRM does not change much and the intensity de-creases gradually and smoothly during ac cleaning, then the remanent magnetization is stable and chiefly composed of TRM. This argument is extended as a consistency check to detect unwanted effects during laboratory heating. A simple procedure which employs orientated samples and a short heating (15 min) for TRM acquisition in the laboratory has been used for determining the ancient geomagnetic field intensity using seven volcanic rocks of Late Cenozoic age from central Mexico. The main reliability tests are based on the stability of direction, the close correspondence of the entire coercitivity spectra of both NRM and TRM to ac demagnetization, the low scatter of TRM directions, close correspondence of the TRM directions and the direction of the laboratory magnetic field, proportionality of TRM intensities to applied field, susceptibility comparison before and after heating, and the within-unit consistency of palaeointensity determinations. 相似文献
16.
《Basin Research》2018,30(4):766-782
This paper proposes a new methodology to improve the location of potential karstified areas by gravity inversion of a 3D geological model. A geological 3D model is built from surface observations, 2D seismic reflection profiles and well data. The reliability of this geological 3D model obtained from integration, interpretation and interpolation of such data is first tested against the structural consistency of the model. Its theoretical gravimetric response is compared to gravity field during the forward problem in order to evaluate the validity/robustness of the geological model. The coherency between the gravity field and the gravimetric response is tested. The litho‐inversion modelling quantifies the distribution of rock density in a probabilistic way, taking into account the geology and physical properties of rocks, while respecting the geological structures represented in the 3D model. The result of the inversion process provides a density distribution within carbonate formations that can be discussed in term of karstification distribution. Thus, lower densities correlate with areas that are strongly karstified. Conversely, higher than mean densities are found in carbonate formations mostly located under marly and impervious formations, preserving carbonate from karstification and paleokarstification. 相似文献
17.
18.
基于粗集、遗传神经网络的环境质量评价方法利用粗集对属性的归约功能将数据库中的数据进行归约，并将归约后的数据作为训练数据提供给BP神经网络；再用遗传算法和BP算法相结合的混合算法来训练网络预测模型的结构（在得到最优网络结构的同时也得到网络的最优权值和阈值）。通过粗糙集归约，提高了训练数据表达的清晰度，也减小了BP神经网络的规模，同时利用BP神经网络又克服了粗糙集对噪声数据敏感的影响。这一算法克服了BP算法收敛速度慢、易陷入局部极小等缺陷，实例证明提高了预测精度。 相似文献
19.
In this paper, a variation series of snow cover and seasonal freeze-thaw layer from 1965 to 2004 on the Tibetan Plateau has been established by using the observation data from meteorological stations. The sliding T-test, M-K test and B-G algorithm are used to verify abrupt changes of snow cover and seasonal freeze-thaw layer in the Tibetan plateau. The results show that the snow cover has not undergone an abrupt change, but the seasonal freeze-thaw layer obviously witnessed a rapid degradation in 1987, with the frozen soil depth being reduced by about 15 cm. It is also found that when there is less snow in the plateau region, precipitation in South China and Southwest China increases. But when the frozen soil is deep, precipitation in most of China apparently decreases. Both snow cover and seasonal freeze-thaw layer on the plateau can be used to predict the summer precipitation in China. However, if the impacts of snow cover and seasonal freeze-thaw layer are used at the same time, the predictability of summer precipitation can be significantly improved. The significant correlation zone of snow is located in middle reaches of the Yangtze River covering the Hexi Corridor and northeastern Inner Mongolia, and the seasonal freeze-thaw layer exists in Mt. Nanling, northern Shannxi and northwestern part of North China. The significant correlation zone of simultaneous impacts of snow cover and seasonal freeze-thaw layer is larger than that of either snow cover or seasonal freeze-thaw layer. There are three significant correlation zones extending from north to south: the north zone spreads from Mt. Daxinganling to the Hexi Corridor, crossing northern Mt. Taihang and northern Shannxi; the central zone covers middle and lower reaches of the Yangtze River; and the south zone extends from Mt. Wuyi to Yunnan and Guizhou Plateau through Mt. Nanling. 相似文献
20.
In order to determine the changing rule of long-term frozen soil strength and elucidate the connection between long-term strength and soil physical properties,frozen loess was subjected to 4,6,8,10,and 50 freeze-thaw cycles,under closed-state conditions in a constant-temperature box.The frozen samples were tested on a spherical template indenter,and the results show that under the effect of repeated freeze-thaw cycles,the long-term strength of frozen loess decreased; changes in the mechanical property indices were highly unstable during the first 10 cycles; the soil strength and density were the greatest at the eighth cycle while the void ratio was the smallest; and after eight cycles all of the indices had less fluctuation and certain rising or falling tendencies.By converting the number of freeze-thaw cycles into elapsed time in the tests,three different forecasting methods of long-term soil strength could be assessed and the soil equivalent cohesive force after 10 years,20 years,or 30 years could be estimated. 相似文献