Landslide susceptibility mapping at Vaz Watershed (Iran) using an artificial neural network model: a comparison between multilayer perceptron (MLP) and radial basic function (RBF) algorithms

Landslide susceptibility and hazard assessments are the most important steps in landslide risk mapping. The main objective of this study was to investigate and compare the results of two artificial neural network (ANN) algorithms, i.e., multilayer perceptron (MLP) and radial basic function (RBF) for spatial prediction of landslide susceptibility in Vaz Watershed, Iran. At first, landslide locations were identified by aerial photographs and field surveys, and a total of 136 landside locations were constructed from various sources. Then the landslide inventory map was randomly split into a training dataset 70 % (95 landslide locations) for training the ANN model and the remaining 30 % (41 landslides locations) was used for validation purpose. Nine landslide conditioning factors such as slope, slope aspect, altitude, land use, lithology, distance from rivers, distance from roads, distance from faults, and rainfall were constructed in geographical information system. In this study, both MLP and RBF algorithms were used in artificial neural network model. The results showed that MLP with Broyden–Fletcher–Goldfarb–Shanno learning algorithm is more efficient than RBF in landslide susceptibility mapping for the study area. Finally the landslide susceptibility maps were validated using the validation data (i.e., 30 % landslide location data that was not used during the model construction) using area under the curve (AUC) method. The success rate curve showed that the area under the curve for RBF and MLP was 0.9085 (90.85 %) and 0.9193 (91.93 %) accuracy, respectively. Similarly, the validation result showed that the area under the curve for MLP and RBF models were 0.881 (88.1 %) and 0.8724 (87.24 %), respectively. The results of this study showed that landslide susceptibility mapping in the Vaz Watershed of Iran using the ANN approach is viable and can be used for land use planning.     

Comparison of machine learning methods for copper ore grade estimation

In this study, machine learning methods such as neural networks, random forests, and Gaussian processes are applied to the estimation of copper grade in a mineral deposit. The performance of these methods is compared to geostatistical techniques, such as ordinary kriging and indicator kriging. To ensure that these comparisons are realistic and relevant, the predictive accuracy is estimated on test instances located in drill holes that are different from the training data. The results of an extensive empirical study in the Sarcheshmeh porphyry copper deposit in Southeastern Iran illustrate that specially designed Gaussian processes with a symmetric standardization of the spatial location inputs and an anisotropic kernel yield the most accurate predictions. Furthermore, significant improvements are obtained when, besides location, information on the rock type is included in the set of predictor variables. This observation highlights the importance of carrying out detailed studies of the geological composition of the deposit to obtain more accurate ore grade predictions.     

Spatial prediction models for shallow landslide hazards: a comparative assessment of the efficacy of support vector machines,artificial neural networks,kernel logistic regression,and logistic model tree

Preparation of landslide susceptibility maps is considered as the first important step in landslide risk assessments, but these maps are accepted as an end product that can be used for land use planning. The main objective of this study is to explore some new state-of-the-art sophisticated machine learning techniques and introduce a framework for training and validation of shallow landslide susceptibility models by using the latest statistical methods. The Son La hydropower basin (Vietnam) was selected as a case study. First, a landslide inventory map was constructed using the historical landslide locations from two national projects in Vietnam. A total of 12 landslide conditioning factors were then constructed from various data sources. Landslide locations were randomly split into a ratio of 70:30 for training and validating the models. To choose the best subset of conditioning factors, predictive ability of the factors were assessed using the Information Gain Ratio with 10-fold cross-validation technique. Factors with null predictive ability were removed to optimize the models. Subsequently, five landslide models were built using support vector machines (SVM), multi-layer perceptron neural networks (MLP Neural Nets), radial basis function neural networks (RBF Neural Nets), kernel logistic regression (KLR), and logistic model trees (LMT). The resulting models were validated and compared using the receive operating characteristic (ROC), Kappa index, and several statistical evaluation measures. Additionally, Friedman and Wilcoxon signed-rank tests were applied to confirm significant statistical differences among the five machine learning models employed in this study. Overall, the MLP Neural Nets model has the highest prediction capability (90.2 %), followed by the SVM model (88.7 %) and the KLR model (87.9 %), the RBF Neural Nets model (87.1 %), and the LMT model (86.1 %). Results revealed that both the KLR and the LMT models showed promising methods for shallow landslide susceptibility mapping. The result from this study demonstrates the benefit of selecting the optimal machine learning techniques with proper conditioning selection method in shallow landslide susceptibility mapping.     

合肥义城地区土壤重金属污染评价中典型插值方法的对比

空间插值对于土壤中重金属元素的空间分布及污染评价具有重要意义。对合肥义城地区土壤中的Cu、Pb、Zn、Cd、As、Hg等污染重金属元素,以常用且具有代表性的反距离加权法、径向基函数法、普通克里格法,进行了空间插值的对比验证分析和评价。通过对各种元素的空间插值各种误差进行综合比较的结果表明：Cu、Pb、As元素采用普通克里格法进行插值结果最优,而Zn元素采用反距离加权法最优,对于Cd、Hg元素则径向基函数插值法最优。     

Estimation of radon as an earthquake precursor: A neural network approach

An artificial neural networks (ANN) approach combined with Fourier Transform based selection of time period in the time series Radon Emission Data has been presented and shown to improve event prediction rates and reduce false alarms in Earthquake Event Identification over the traditional multiple linear regression techniques. The paper presents a neural networks system using radial basis function (RBF) network as an alternative to traditional statistical regression technique in isolating Radon Emission Anomaly caused by seismic activities. The RBF model has been developed to accept and predict earthquakes events based on a known data set of Radon Emanation, Metrological parameters and actual earthquake events. Subsequently, the model was tested and evaluated on a future data set and a prediction rate of 87.8%, if a reduced false alarm was achieved, the results obtained are better than the traditional techniques.     

A comparative analysis among computational intelligence techniques for dissolved oxygen prediction in Delaware River

Most of the water quality models previously developed and used in dissolved oxygen (DO) prediction are complex. Moreover, reliable data available to develop/calibrate new DO models is scarce. Therefore, there is a need to study and develop models that can handle easily measurable parameters of a particular site, even with short length. In recent decades, computational intelligence techniques, as effective approaches for predicting complicated and significant indicator of the state of aquatic ecosystems such as DO, have created a great change in predictions. In this study, three different AI methods comprising: (1) two types of artificial neural networks (ANN) namely multi linear perceptron (MLP) and radial based function (RBF); (2) an advancement of genetic programming namely linear genetic programming (LGP); and (3) a support vector machine (SVM) technique were used for DO prediction in Delaware River located at Trenton, USA. For evaluating the performance of the proposed models, root mean square error (RMSE), Nash–Sutcliffe efficiency coefficient (NS), mean absolute relative error (MARE) and, correlation coefficient statistics (R) were used to choose the best predictive model. The comparison of estimation accuracies of various intelligence models illustrated that the SVM was able to develop the most accurate model in DO estimation in comparison to other models. Also, it was found that the LGP model performs better than the both ANNs models. For example, the determination coefficient was 0.99 for the best SVM model, while it was 0.96, 0.91 and 0.81 for the best LGP, MLP and RBF models, respectively. In general, the results indicated that an SVM model could be employed satisfactorily in DO estimation.     

Comparison of ordinary and disjunctive kriging in an Indian iron ore deposit

This paper presents the results of disjunctive kriging applied to a supergene iron ore deposit of Bailadila Range of India. Disjunctive kriging is applied firstly to compare estimates of the blocks by ordinary kriging and secondly to estimate benchwise local recoverable reserves of the orebody. Good agreement exists between block estimates by ordinary kriging and disjunctive kriging except for peripheral blocks with less borehole information. Estimation of benchwise reserves shows that the behavior of the distribution of grades is different in various benches. The study shows that disjunctive kriging can be applied successfully for estimation of local recoverable reserves in the case of a good grade hematite iron ore deposit.     

Application of neural network and ANFIS model for earthquake occurrence in Iran

This study examined the spatial-temporal variations in seismicity parameters for the September 10^th, 2008 Qeshm earthquake in south Iran. To this aim, artificial neural networks and Adaptive Neural Fuzzy Inference System (ANFIS) were applied. The supervised Radial Basis Function (RBF) network and ANFIS model were implemented because they have shown the efficiency in classification and prediction problems. The eight seismicity parameters were calculated to analyze spatial and temporal seismicity pattern. The data preprocessing that included normalization and Principal Component Analysis (PCA) techniques was led before the data was fed into the RBF network and ANFIS model. Although the accuracy of RBF network and ANFIS model could be evaluated rather similar, the RBF exhibited a higher performance than the ANFIS for prediction of the epicenter area and time of occurrence of the 2008 Qeshm main shock. A proper training on the basis of RBF network and ANFIS model might adopt the physical understanding between seismic data and generate more effective results than conventional prediction approaches. The results of the present study indicated that the RBF neural networks and the ANFIS models could be suitable tools for accurate prediction of epicenteral area as well as time of occurrence of forthcoming strong earthquakes in active seismogenic areas.     

System reliability analysis of slopes using multilayer perceptron and radial basis function networks

This paper presents a system reliability analysis method for soil slopes on the basis of artificial neural networks with computer experiments. Two types of artificial neural networks, multilayer perceptrop (MLP) and radial basis function networks (RBFNs), are tested on the studied problems. Computer experiments are adopted to generate samples for constructing the response surfaces. On the basis of the samples, MLP and RBFN are used for establishing the response surface to approximate the limit state function, and Monte Carlo simulation is performed via the MLP and RBFN response surfaces to estimate the system failure probability of slopes. Experimental results on 3 examples show the effectiveness of the proposed methodology.     

Application of disjunctive kriging for estimating economic grade distribution in an iron ore deposit: A case study of the Choghart North Anomaly,Iran

Most significant iron ore deposits in Iran are located in Central Iran Zone. These deposits belong to the Bafq mining district. The Bafq mining district is located in the Early Cambrian Kashmar-Kerman volcanic arc of Central Iran. Linear estimation of regionalized variables (for example by inverse distance weighting or ordinary Kriging) results in relatively high estimation variances, i.e. the estimates have very low precision. Assessment of project economics (or other critical decision making) based on linear estimation is therefore risky. Non-linear estimation methods like disjunctive kriging perform better and the lower estimation variance allows less risky economic decision-making. Another advantage of disjunctive kriging is that it allows estimation of functions of the primary variable, which here is the grade (Fe %) of the ore. In particular it permits estimation of indicator functions defined using thresholds on the primary variable. This paper is devoted to application of disjunctive kriging method in Choghart North Anomaly iron ore deposit in Central Iran, Yazd province, Iran. In this study, the Fe concentration of Choghart North Anomaly iron ore deposit was modelled and estimated. The exploration data consists of borehole samples measuring the Fe concentration. A Gaussian isofactorial model is fitted to these data and disjunctive kriging was used to estimate the regionalized variable (Fe %) at unsampled locations and to assess the probabilities that the actual concentrations exceed a threshold value at a given location. Consequently a three dimensional model of probability of exceeding a threshold value and the estimated value are provided by disjunctive kriging to divide the ore into an economic and uneconomic part on the basis of estimation of indicator functions using thresholds grades defined on point support. The tools and concepts are complemented by a set of computer programs that are applied to the case study. The study showed that disjunctive kriging can be applied successfully for modeling the grade of an ore deposit. Results showed that the correlation between the estimated value and real value at locations close to each other is 81.9%.     

Evaluating Thickness of Bauxite Deposit Using Indicator Geostatistics and Fuzzy Estimation

Evaluating the geological properties of a mineral deposit is a fundamental task for mine planning and it requires an assessment of reserve parameters such as thickness and grade. This paper presents a linguistic model for estimating bauxite thickness within a deposit in a fuzzy environment using indicator geostatistics and fuzzy modeling. The proposed model consists of two main stages: determining the orebody boundary and estimating the thickness. In order to estimate the thickness, a rule‐based fuzzy inference mechanism has been developed based on data statistics. Results and performance of the model have been compared with that of a well‐known conventional technique, geostatistics (kriging), and it is shown that the proposed model has bigger estimation power. In addition, the fuzzy approach is more flexible than the kriging approach. The fuzzy methodology used in the present paper is convenient for modeling reserve parameters.     

A comparative study between sequential Gaussian simulation and kriging method grade modeling in open-pit mining

Grade estimation is very important in designing open pits. In the process of grade estimation, underestimation can result in loss of economic ore, whereas overestimation would unnecessarily increase stripping ratio. Normally, kriging method, which suffers from underestimation and/or overestimation due to smoothing effect, is used for grade estimation. To overcome drawbacks of the kriging method, more efficient techniques such as conditional simulation can be applied. In this paper, utilizing sequential Gaussian conditional simulation, grade models were constructed for Sungun copper deposit situated in the North West of Iran. According to the obtained results, it was observed that conditional simulation can effectively cope with the weakness of kriging method. Also, it was observed that as compared to the kriging method, grade distribution, resulted from the conditional simulation, is almost identical to that of the real exploration data. Accordingly, using conditional simulation, the amount of mineable ore was significantly increased, and also, average net present value as the mines’ most important economic indicator was improved by 40%.     

Comparison between different kriging estimators

Six different geostatistical estimators (linear kriging, lognormal kriging, and disjunctive kriging, each with and without a nonbias, i.e., universality condition) were compared using data from a polymetallic deposit in Algeria. The differences between estimators with and without the nonbias condition were far more pronounced than between the different kriging methods. This highlights the importance of choosing an appropriate stationarity model for the data. The criterion concerning kriging weight of the mean in simple kriging, proposed by Remacre (1984, 1987) and Rivoirard (1984) was found to be helpful for determining blocks where the choice of the stationarity hypothesis was critical.     

Using geology to control geostatistics in the Hellyer deposit

The Hellyer orebody, a polymetallic massive sulfide deposit, was discovered in western Tasmania by Aberfoyle in 1983. Delineation diamond drilling was carried out in 1984 on a nominal 50-m square grid pattern to outline the resource. Resource estimation methods were influenced by the requirement to develop a regular block model for conceptual mine planning studies. Detailed geological interpretation indicates that the interpolation technique must take into account several important features to retain geological credibility. The deposit has sharp limits defined by visual geological contacts with virtually barren enclosing rocks. Lateral terminations are rapid with no interfingering internal waste. The dip and strike are variable and a major fault with a measurable displacement cuts acutely through the center of the deposit. Ore grades are reasonably correlateable within specific layers from hole to hole indicating a significant across-dip anisotropy. A hanging wall enriched zone is well-defined throughout the deposit. To overcome the variable geometry problems, a stratigraphic coordinate system was defined arbitrarily to replace the normal z coordinate. This allowed variography in stratigraphic layers. Blocks to be estimated were constrained by hand-drawn and subsequently digitized hanging wall and footwall contours. Each block was ascribed a stratigraphic coordinate by calculating its spatial position in relation to nearby stratigraphic unit boundaries within the massive sulfide body. Estimates were generated for each element by ordinary linear kriging. Despite the relatively sparse data in a large massive deposit, the customized technique developed for Hellyer has provided a reliable model of spatial grade distribution by combining conventional geostatistical methods with careful geological observation and interpretation. Some geometry problems remain which are the subject of ongoing studies.This paper was presented at MGUS 87 Conference, Redwood City, California, 14 April 1987.     

Fuzzy expert system for economic zonation of an ornamental slate deposit

Investigation of deposits for traditional extraction activities (metals and coal) has generally been based on determining grade, or content, of the required material. In order to apply the grade concept to an ornamental rock such as slate, it is first necessary to define the variables that determine both the geotechnical recovery rate for the rock mass — which conditions the size of the extracted blocks – and the aesthetic features of the slate — which define the quality of the slabs as potential roofing material.

For this research, geotechnical and aesthetic data for a slate deposit were collected from 16 continuous core borehole samples. A fuzzy expert system was then developed using this data, defining the rock mass recovery rate and slab quality in accordance with the criteria of a slate expert, producing as a final output a zonation of the deposit in terms of top quality slate, medium quality slate or waste.

A mathematical model based on fuzzy logic was chosen due to the fact that the boundaries between different quality groups in a deposit are not clearly distinguished. Moreover, quality also depends on a company's infrastructures for transformation of the blocks, and also on its commercial strategies.     

克里格法在离子吸附型稀土矿勘查储量估算中的应用

我国在离子吸附型稀土矿勘查工作中,一般采用地质块段法估算储量,块段法是将矿体划分为不同厚度的块段投影到水平或垂直方向上,块段的划分、各块段的面积和厚度、品位都会影响储量估算结果。本文以赣南某离子吸附型稀土矿床作为研究对象,基于先期勘探钻孔数据资料,运用三维建模软件创建了该矿床钻孔数据库,建立了矿区内矿体的三维DTM模型;采用克里格法对矿体进行稀土氧化物品位分析,将克里格法的储量计算结果与块段法的储量计算结果作对比分析。结果显示,克里格法计算的矿体体积比块段法增加了11.8%,稀土氧化物储量增加了15%,与实际勘探数据相比较,克里格法的计算结果基本合理,且具有快速、准确、方便的特点。本文利用自主开发的以克里格法为基础的三维数字矿山经济评价系统中价格-边界品位敏感性分析模块,动态设置边界品位,灵活圈定不同价格下经济可采的矿体边界,如当精矿的市场价格从10万元/吨变化为12万元/吨时,通过计算获得了此矿山经济可采矿体的空间扩展范围。基于克里格法的三维估算系统能够帮助矿山选择合理的采矿工程布置,有利于满足矿山动态管理的需要以及保证矿产资源的合理利用。     

Application of universal kriging for estimation of earthquake ground motion: Statistical significance of results

Universal kriging is compared with ordinary kriging for estimation of earthquake ground motion. Ordinary kriging is based on a stationary random function model; universal kriging is based on a nonstationary random function model representing first-order drift. Accuracy of universal kriging is compared with that for ordinary kriging; cross-validation is used as the basis for comparison. Hypothesis testing on these results shows that accuracy obtained using universal kriging is not significantly different from accuracy obtained using ordinary kriging. Tests based on normal distribution assumptions are applied to errors measured in the cross-validation procedure;t andF tests reveal no evidence to suggest universal and ordinary kriging are different for estimation of earthquake ground motion. Nonparametric hypothesis tests applied to these errors and jackknife statistics yield the same conclusion: universal and ordinary kriging are not significantly different for this application as determined by a cross-validation procedure. These results are based on application to four independent data sets (four different seismic events).