Development of discharge-stage curves affected by hysteresis using time varying models,model trees and neural networks

Flow data forms the base on which much of the edifice of water management is raised. However, flow measurements are expensive and difficult to conduct. Therefore, the more accessible stage measurements are employed in combination with stage–discharge relationships. Setting up such relationships is often infeasible using traditional regression techniques. Two case studies are examined that show hystereses using various approaches, namely (1) single rating curves, (2) rating curves with dynamic correction, (3) artificial neural networks (ANN) and (4) M5′ model trees. All methods outperform the traditional rating curve. The presented approach that uses a dynamically corrected rating curve delivers accurate results and allows for physical interpretation. The ANNs mimic the calibration data precisely, but suffer from overfitting when a small amount of data is applied for training. The rarely used M5′ model tree's architecture is easier to interpret than that of neural networks and delivers more accurate results.     

A critical review of the most popular types of neuro control

In this review article, the most popular types of neural network control systems are briefly introduced and their main features are reviewed. Neuro control systems are defined as control systems in which at least one artificial neural network (ANN) is directly involved in generating the control command. Initially, neural networks were mostly used to model system dynamics inversely to produce a control command which pushes the system towards a desired or reference value of the output (1989). At the next stage, neural networks were trained to track a reference model, and ANN model reference control appeared (1990). In that method, ANNs were used to extend the application of adaptive reference model control, which was a well‐known control technique. This attitude towards the extension of the application of well‐known control methods using ANNs was followed by the development of ANN model‐predictive (1991), ANN sliding mode (1994) and ANN feedback linearization (1995) techniques. As the first category of neuro controllers, inverse dynamics ANN controllers were frequently used to form a control system together with other controllers, but this attitude faded as other types of ANN control systems were developed. However, recently, this approach has been revived. In the last decade, control system designers started to use ANNs to compensate/cancel undesired or uncertain parts of systems' dynamics to facilitate the use of well‐known conventional control systems. The resultant control system usually includes two or three controllers. In this paper, applications of different ANN control systems are also addressed. Copyright © 2011 John Wiley and Sons Asia Pte Ltd and Chinese Automatic Control Society     

Comparison of artificial neural networks and support vector machine classifiers for land cover classification in Northern China using a SPOT-5 HRG image

This article presents a sufficient comparison of two types of advanced non-parametric classifiers implemented in remote sensing for land cover classification. A SPOT-5 HRG image of Yanqing County, Beijing, China, was used, in which agriculture and forest dominate land use. Artificial neural networks (ANNs), including the adaptive backpropagation (ABP) algorithm, Levenberg–Marquardt (LM) algorithm, Quasi-Newton (QN) algorithm and radial basis function (RBF) were carefully tested. The LM–ANN and RBF–ANN, which outperform the other two, were selected to make a detailed comparison with support vector machines (SVMs). The experiments show that those well-trained ANNs and SVMs have no significant difference in classification accuracy, but the SVM usually performs slightly better. Analysis of the effect of the training set size highlights that the SVM classifier has great tolerance on a small training set and avoids the problem of insufficient training of ANN classifiers. The testing also illustrates that the ANNs and SVMs can vary greatly with regard to training time. The LM–ANN can converge very quickly but not in a stable manner. By contrast, the training of RBF–ANN and SVM classifiers is fast and can be repeatable.     

An artificial neural network approach for landslide hazard zonation in the Bhagirathi (Ganga) Valley,Himalayas

Landslides are natural hazards that cause havoc to both property and life every year, especially in the Himalayas. Landslide hazard zonation (LHZ) of areas affected by landslides therefore is essential for future developmental planning and organization of various disaster mitigation programmes. The conventional Geographical Information System (GIS)-based approaches for LHZ suffer from the subjective weight rating system where weights are assigned to different causative factors responsible for triggering a landslide. Alternatively, artificial neural networks (ANNs) may be applied. These are considered to be independent of any strict assumptions or bias, and they determine the weights objectively in an iterative fashion. In this study, an ANN has been applied to generate an LHZ map of an area in the Bhagirathi Valley, Himalayas, using spatial data prepared from IRS-1B satellite sensor data and maps from other sources. The accuracy of the LHZ map produced by the ANN is around 80% with a very small training dataset. The distribution of landslide hazard zones derived from ANN shows similar trends as that observed with the existing landslides locations in the field. A comparison of the results with an earlier produced GIS-based LHZ map of the same area by the authors (using the ordinal weight rating method) indicates that ANN results are better than the earlier method.     

Using a heuristic approach to derive a grey-box model through an artificial neural network knowledge extraction technique

Artificial neural networks (ANNs) are primarily used in academia for their ability to model complex nonlinear systems. Though ANNs have been used to solve practical problems in industry, they are not typically used in nonacademic environments because they are not very well understood, complicated to implement, or have the reputation of being a “black-box” model. Few mathematical models exist that outperform ANNs. If a highly accurate model can be constructed, the knowledge should be used to understand and explain relationships in a system. Output surfaces can be analyzed in order to gain additional knowledge about a system being modeled. This paper presents a systematic approach to derive a “grey-box” model from the knowledge obtained from the ANN. A database for an automobile’s gas mileage performance is used as a case study for the proposed methodology. The results show a greater ability to generalize system behavior than other benchmarked methods.     

Comparison of country risk models: hybrid neural networks,logit models,discriminant analysis and cluster techniques

This paper looks at the ability of a relatively new technique, hybrid ANN's, to predict country risk rating. These models are compared with traditional statistical techniques and conventional ANN models. The performance of hierarchical cluster analysis and another type of ANN, the self-organizing map were also investigated, as possible methods for making country risk analysis with visual effects. The results indicate that hybrid neural networks outperform all other models. This suggests that for researchers, policymakers and others interested in early warning systems, hybrid network may be a useful tool for country risk analysis.     

An unsupervised artificial neural network versus a rule-based approach for fault detection and identification in an automated assembly machine

Artificial neural networks (ANNs) are suitable for fault detection and identification (FDI) applications because of their pattern recognition abilities. In this study, an unsupervised ANN based on Adaptive Resonance Theory (ART) is tested for FDI on an automated O-ring assembly machine testbed, and its performance and practicality are compared to a conventional rule-based method. Three greyscale sensors and two redundant limit switches are used as cost-effective sensors to monitor the machine’s assembly process. Sensor data are collected while the machine is operated under normal condition, as well as 10 fault conditions. Features are selected from the raw sensor data, and data sets are created for training and testing the ANN. The performance of the ANN for detecting and identifying known, unknown and multiple faults is evaluated; the performance is compared to a conventional rule-based method using the same data sets. Results show that the ART ANN is able to achieve excellent fault detection performance with minimal modeling requirements; however, the performance depends on careful tuning of its vigilance parameter. Although the rule-based system requires more effort to set up, it is judged to be more useful when unknown or multiple faults are present. The ART network creates new outputs for unknown and multiple fault conditions, but it does not give any more information as to what the new fault is. By contrast, the rule-based method is able to generate symptoms that clearly identify the unknown and multiple fault conditions. Thus, the rule-based method is judged to be the most feasible method for FDI applications.     

Probabilistic neural networks for the identification of qualified audit opinions

Prior studies that examine the application of neural networks in auditing investigate the efficiency of artificial neural networks (ANNs). In the present study, considering the well known disadvantages of artificial neural network, we propose the application of probabilistic neural networks (PNNs) that combine the computational power and flexibility of ANNs, while managing to retain simplicity and transparency. The sample consists of 264 financial statements that received a qualified audit opinion over the period 1997–2004 and 3069 unqualified ones, from 881 firms listed on the London Stock Exchange. The results demonstrate the high explanatory power of the PNN model in explaining qualifications in audit reports. The model is also found to outperform traditional ANN models, as well as logistic regression. Sensitivity analysis is used to assess the relative importance of the input variables and to analyze their role in the auditing process.     

ANN modeling of DNA sequences: new strategies using DNA shape code

Two new encoding strategies, namely, wedge and twist codes, which are based on the DNA helical parameters, are introduced to represent DNA sequences in artificial neural network (ANN)-based modeling of biological systems. The performance of the new coding strategies has been evaluated by conducting three case studies involving mapping (modeling) and classification applications of ANNs. The proposed coding schemes have been compared rigorously and shown to outperform the existing coding strategies especially in situations wherein limited data are available for building the ANN models.     

Mapping tropical forest structure in southeastern Madagascar using remote sensing and artificial neural networks

Tropical forest condition has important implications for biodiversity, climate change and human needs. Structural features of forests can serve as useful indicators of forest condition and have the potential to be assessed with remotely sensed imagery, which can provide quantitative information on forest ecosystems at high temporal and spatial resolutions. Herein, we investigate the utility of remote sensing for assessing, predicting and mapping two important forest structural features, stem density and basal area, in tropical, littoral forests in southeastern Madagascar. We analysed the relationships of basal area and stem density measurements to the Normalised Difference Vegetation Index (NDVI) and radiance measurements in bands 3, 4, 5 and 7 from the Landsat Enhanced Thematic Mapper Plus (ETM+). Strong relationships were identified among all of the individual bands and field based measurements of basal area (p<0.01) while there were weak and insignificant relationships among spectral response and stem density measurements. NDVI was not significantly correlated with basal area but was strongly and significantly correlated with stem density (r=−0.69, p<0.01) when using a subset of the data, which represented extreme values. We used an artificial neural network (ANN) to predict basal area from radiance values in bands 3, 4, 5 and 7 and to produce a predictive map of basal area for the entire forest landscape. The ANNs produced strong and significant relationships between predicted and actual measures of basal area using a jackknife method (r=0.79, p<0.01) and when using a larger data set (r=0.82, p<0.01). The map of predicted basal area produced by the ANN was assessed in relation to a pre-existing map of forest condition derived from a semi-quantitative field assessment. The predictive map of basal area provided finer detail on stand structural heterogeneity, captured known climatic influences on forest structure and displayed trends of basal area associated with degree of human accessibility. These findings demonstrate the utility of ANNs for integrating satellite data from the Landsat ETM+ spectral bands 3, 4, 5 and 7 with limited field survey data to assess patterns in basal area at the landscape scale.     

基于人工神经网络组合预测油田产量

油田原油产量的准确预测可以对油田的生产管理进行合理的指导。该文探讨了应用神经网络组合方法预测油田产量，对开井数、含水率、动用储量以及往年产量同未来产量之间的复杂关系建立模型。采用了两层预测系统：第一层包含两个神经网络，一个多层前馈网络和一个函数链接网络；第二层是把第一层的两个网络输出进行组合。研究了五种不同的组合算法：平均法、最小平方回归法、模糊逻辑法、自适应前馈神经网络法和自适应函数链接神经网络法。根据油品类型分为稀油、热采稠油、常规稠油和总产量四组数据，对上述方法进行了测试，结果表明应用人工神经网络的组合预测方法优于其他的预测方法，而且适用范围广。     

TACO-miner: An ant colony based algorithm for rule extraction from trained neural networks

Extracting classification rules from data is an important task of data mining and gaining considerable more attention in recent years. In this paper, a new meta-heuristic algorithm which is called as TACO-miner is proposed for rule extraction from artificial neural networks (ANN). The proposed rule extraction algorithm actually works on the trained ANNs in order to discover the hidden knowledge which is available in the form of connection weights within ANN structure. The proposed algorithm is mainly based on a meta-heuristic which is known as touring ant colony optimization (TACO) and consists of two-step hierarchical structure. The proposed algorithm is experimentally evaluated on six binary and n-ary classification benchmark data sets. Results of the comparative study show that TACO-miner is able to discover accurate and concise classification rules.     

Fuzzy learning vector quantization for hyperspectral coastal vegetation classification

Artificial neural networks (ANNs) may be of significant value in extracting vegetation type information in complex vegetation mapping problems, particularly in coastal wetland environments. Unsupervised, self-organizing ANNs have not been employed as frequently as supervised ANNs for vegetation mapping tasks, and further remote sensing research involving fuzzy ANNs is also needed. In this research, the utility of a fuzzy unsupervised ANN, specifically a fuzzy learning vector quantization (FLVQ) ANN, was investigated in the context of hyperspectral AVIRIS image classification. One key feature of the neural approach is that unlike conventional hyperspectral data processing methods, endmembers for a given scene, which can be difficult to determine with confidence, are not required for neural analysis. The classification accuracy of FLVQ was comparable to a conventional supervised multi-layer perceptron, trained with backpropagation (MLP) (KHAT (Kˆ) accuracy: 82.82% and 84.66%, respectively; normalized accuracy: 74.60% and 75.85%, respectively), with no significant difference at the 95% confidence level. All neural algorithms in the experiment yielded significantly higher classification accuracies than the conventional endmember-based hyperspectral mapping method assessed (i.e., matched filtering, where Kˆ accuracy = 61.00% and normalized accuracy = 57.96%). FLVQ was also dramatically more computationally efficient than the baseline supervised and unsupervised ANN algorithms tested, including the MLP and the Kohonen self-organizing map (SOM), respectively. The 400-neuron FLVQ network required only 3.6% of the computation time used by the MLP network, and only 5.9% of the MLP time was used by the 588-neuron FLVQ network. In addition, the 400-neuron FLVQ used only 16.7% of the time used by the 400-neuron SOM for model development.     

Enhancement of artificial neural network learning using centripetal accelerated particle swarm optimization for medical diseases diagnosis

In recent decades, artificial neural networks (ANNs) have been extensively applied in different areas such as engineering, medicine, business, education, manufacturing and so on. Nowadays, ANNs are as a hot research in medicine especially in the fields of medical disease diagnosis. To have a high efficiency in ANN, selection of an appropriate architecture and learning algorithm is very important. ANN learning is a complex task and an efficient learning algorithm has a significant role to enhance ANN performance. In this paper, a new meta-heuristic algorithm, centripetal accelerated particle swarm optimization (CAPSO), is applied to evolve the ANN learning and accuracy. The algorithm is based on an improved scheme of particle swarm algorithm and Newton’s laws of motion. The hybrid learning of CAPSO and multi-layer perceptron (MLP) network, CAPSO-MLP, is used to classify the data of nine standard medical datasets of Hepatitis, Heart Disease, Pima Indian Diabetes, Wisconsin Prognostic Breast Cancer, Parkinson’s disease, Echocardiogram, Liver Disorders, Laryngeal 1 and Acute Inflammations. The performance of CAPSO-MLP is compared with those of PSO, gravitational search algorithm and imperialist competitive algorithm on MLP. The efficiency of methods are evaluated based on mean square error, accuracy, sensitivity, specificity, area under the receiver operating characteristics curve and statistical tests of \(t\) -test and Wilcoxon’s signed ranks test. The results indicate that CAPSO-MLP provides more effective performance than the others for medical disease diagnosis especially in term of unseen data (testing data) and datasets with high missing data values.     

Using online cognitive tasks to predict mathematics low school achievement

This study aimed to analyze the validity of an online cognitive screening battery to predict mathematic school achievement using artificial neural networks (ANNs). The tasks were designed to measure; selective attention, visuo-spatial working memory, mental rotation, and arithmetic ability in an online, game-like format. In the first study, we investigated the cognitive performance of students with low and typical achievement in mathematics and language. In the second study, we developed an ANN to classify mathematics school achievement. Finally, we tested the adequacy of this network to classify an unknown sample to the ANN. Most of the performance differences in the battery were related to mathematics achievement. The ANN was able to predict mathematics achievement with acceptable accuracy and presented equivalent results in a simulation involving a different sample. We suggest that this assessment model combining ANNs and online cognitive tasks may be a valuable tool to research low school achievement in school settings.     

An Improved Particle Swarm Optimization for Evolving Feedforward Artificial Neural Networks

This paper presents a new evolutionary artificial neural network (ANN) algorithm named IPSONet that is based on an improved particle swarm optimization (PSO). The improved PSO employs parameter automation strategy, velocity resetting, and crossover and mutations to significantly improve the performance of the original PSO algorithm in global search and fine-tuning of the solutions. IPSONet uses the improved PSO to address the design problem of feedforward ANN. Unlike most previous studies on only using PSO to evolve weights of ANNs, this study puts its emphasis on using the improved PSO to evolve simultaneously structure and weights of ANNs by a specific individual representation and evolutionary scheme. The performance of IPSONet has been evaluated on several benchmarks. The results demonstrate that IPSONet can produce compact ANNs with good generalization ability.     

The asymptotic expansion and extrapolation of trapezoidal rule for integrals with fractional order singularities

This paper is aimed at deriving the error asymptotic expansion of trapezoidal rule approximation to integrals for the functions with fractional derivatives or algebraic singularities at some points and from which to design a modified Romberg extrapolation algorithm to effectively compute some singular integrals. Firstly, high-order local fractional derivatives are defined, then a general fractional Taylor's expansion is derived. Secondly, the error asymptotic expansion of trapezoidal rule for these integrals is obtained directly by using the formula of sums of non-integer powers and the general fractional Taylor's expansion. This method is different from the previous work. Thirdly, a modified Romberg extrapolation algorithm is designed to get numerical results efficiently. Finally, numerical examples are presented to verify the correctness of the theoretical analysis and the effectiveness of the extrapolation method.     

Estimation of chemical oxygen demand by ultraviolet spectroscopic profiling and artificial neural networks

A simple method based on the mathematical treatment of spectral absorbance profiles in conjunction with artificial neural networks (ANNs) is demonstrated for rapidly estimating chemical oxygen demand (COD) values of wastewater samples. In order to improve spectroscopic analysis and ANN training time as well as to reduce the storage space of the trained ANN algorithm, it is necessary to decrease the ANN input vector size by extracting unique characteristics from the raw input pattern. Key features from the spectral absorbance pattern were therefore selected to obtain the spectral absorbance profile, reducing the ANN input vector from 160 to 10 selected inputs. The results indicate that the COD values obtained from the selected absorbance profiles agreed well with those obtained from the entire absorbance pattern. The spectral absorbance profile technique was also compared to COD values estimated by a multiple linear regression (MLR) model to validate whether ANNs were better and more robust models for rapid COD analysis. It was found that the ANN model predicted COD values closer to standard COD values than the MLR model.     

A hybrid pansharpening approach and multiscale object-based image analysis for mapping diseased pine and oak trees

We developed a multiscale object-based classification method for detecting diseased trees (Japanese Oak Wilt and Japanese Pine Wilt) in high-resolution multispectral satellite imagery. The proposed method involved (1) a hybrid intensity–hue–saturation smoothing filter-based intensity modulation (IHS-SFIM) pansharpening approach to obtain more spatially and spectrally accurate image segments; (2) synthetically oversampling the training data of the ‘Diseased tree’ class using the Synthetic Minority Over-sampling Technique (SMOTE); and (3) using a multiscale object-based image classification approach. Using the proposed method, we were able to map diseased trees in the study area with a user's accuracy of 96.6% and a producer's accuracy of 92.5%. For comparison, the diseased trees were mapped at a user's accuracy of 84.0% and a producer's accuracy of 70.1% when IHS pansharpening was used alone and a single-scale classification approach was implemented without oversampling the ‘Diseased tree’ class.     

The self-trapping attractor neural network. I. Analysis of a simple 1-D model

Attractor neural networks (ANNs) based on the Ising model are naturally fully connected and are homogeneous in structure. These features permit a deep understanding of the underlying mechanism, but limit the applicability of these models to the brain. A more biologically realistic model can be derived from an equally simple physical model by utilizing recurrent self-trapping inputs to supplement very sparse intranetwork interactions. This paper reports the analysis of a one-dimensional (1-D) ANN coupled to a second system that computes overlaps with a single stored memory. Results show that: 1) the 1-D self-trapping model is equivalent to an isolated ANN with both full connectivity of one strength and nearest neighbor synapses of an independent strength; 2) the dynamics of ANN and self-trapping updates are independent; 3) there is a critical synaptic noise level below which memory retrieval occurs; 4) the 1-D self-trapping model converges to a fully connected Hopfield model for zero strength nearest neighbor synapses, and has a greater magnitude memory overlap for nonzero strength nearest neighbor synapses; and (5) the mechanism of self-trapping is an iterative map on the mean overlap as a function of the reentrant input.