Home About us Contact
|
Home About us Contact | ||
|
|
||
Neural Networks (neural + network)
Kinds of Neural Networks Terms modified by Neural Networks Selected AbstractsSTUDY ON THE MOTION ERROR OF MAIN SPINDLE OF LATHE BASED ON THE HARMONIC WAVELET NEURAL NETWORKEXPERIMENTAL TECHNIQUES, Issue 3 2009X.-J. Fu First page of article [source] A NEW APPROACH TO MODELING AND CONTROL OF A FOOD EXTRUSION PROCESS USING ARTIFICIAL NEURAL NETWORK AND AN EXPERT SYSTEMJOURNAL OF FOOD PROCESS ENGINEERING, Issue 1 2001OTILIA POPESCU ABSTRACT The paper presents a new approach to the modeling of the start-up part of a food extrusion process. A neural network model is proposed and its parameters are determined. Simulation results with real data are also presented. The inputs and outputs of the model are among those used by the human operator during the start-up process for control. An intelligent controller structure that uses an expert system and "delta-variations" to modify inputs is also proposed. [source] BP NEURAL NETWORK FOR EVALUATING SENSORY TEXTURE PROPERTIES OF COOKED SAUSAGEJOURNAL OF SENSORY STUDIES, Issue 6 2009QING-LI DONG ABSTRACT In order to replace sensory evaluation by instrumental measurement with more accuracy for texture properties of cooked sausage, correlation analysis between sensory and instrumental texture was established by multiple regression and back propagation (BP) neural network, respectively. Effect of different fat, salt, moisture and starch addition on the texture of cooked sausage was also investigated in this paper. It indicated that the accuracy and goodness of fit of predicting sensory hardness, cohesiveness and juiciness by BP neural network were more significant than those by multiple regressions with lower root mean square error and standard error of prediction. Although both accuracy and bias factors of two models were in acceptable range, BP neural network provides an accurate and selective method for predicting sensory texture evaluation in similar meat products. PRACTICAL APPLICATIONS The effect of different fat, salt, moisture and starch addition on textural properties of cooked sausage could be valuable to the meat industry in order to select the appropriate components for improving the texture of sausage. Artificial neural network technology used in this study can be useful for the fast, on-time and convenient detection of texture measurement by instrumental instead of sensory evaluation. [source] DYNAMIC MODELING OF RETORT PROCESSING USING NEURAL NETWORKSJOURNAL OF FOOD PROCESSING AND PRESERVATION, Issue 2 2002C. R. CHEN ABSTRACT Two neural network approaches , a moving-window and hybrid neural network , which combine neural network with polynomial regression models, were used for modeling F(t) and Qv(t) dynamic functions under constant retort temperature processing. The dynamic functions involved six variables: retort temperature (116,132C), thermal diffusivity (1.5,2.3 × 10,7m2/s), can radius (40,61 mm), can height (40,61 mm), and quality kinetic parameters z (15,39C) and D (150,250 min). A computer simulation designed for process calculations of food thermal processing systems was used to provide the fundamental data for training and generalization of ANN models. Training data and testing data were constructed by both second order central composite design and orthogonal array, respectively. The optimal configurations of ANN models were obtained by varying the number of hidden layers, number of neurons in hidden layer and learning runs, and a combination of learning rules and transfer function. Results demonstrated that both neural network models well described the F(t) and Qv(t) dynamic functions, but moving-window network had better modeling performance than the hybrid ANN models. By comparison of the configuration parameters, moving-window ANN models required more neurons in the hidden layer and more learning runs for training than the hybrid ANN models. [source] PREDICTION OF MECHANICAL PROPERTIES OF CUMIN SEED USING ARTIFICIAL NEURAL NETWORKSJOURNAL OF TEXTURE STUDIES, Issue 1 2010M.H. SAIEDIRAD ABSTRACT In this paper, two artificial neural networks (ANNs) are applied to acquire the relationship between the mechanical properties and moisture content of cumin seed, using the data of quasi-static loading test. In establishing these relationship, the moisture content, seed size, loading rate and seed orientation were taken as the inputs of both models. The force and energy required for fracturing of cumin seed, under quasi-static loading were taken as the outputs of two models. The activation function in the output layer of models obeyed a linear output, whereas the activation function in the hidden layers were in the form of a sigmoid function. Adjusting ANN parameters such as learning rate and number of neurons and hidden layers affected the accuracy of force and energy prediction. Comparison of the predicted and experimented data showed that the ANN models used to predict the relationships of mechanical properties of cumin seed have a good learning precision and good generalization, because the root mean square errors of the predicated data by ANNs were rather low (4.6 and 7.7% for the force and energy, respectively). PRACTICAL APPLICATIONS Cumin seed is generally used as a food additive in the form of powder for imparting flavor to different food preparations and for a variety of medicinal properties. Physical properties of cumin seeds are essential for the design of equipment for handling, harvesting, aeration, drying, storing, grinding and processing. For powder preparation especially the fracture behavior of the seeds are essential. These properties are affected by numerous factors such as size, form and moisture content of the grain and deformation speed. A neural network model was developed that can be used to predict the relationships of mechanical properties. Artificial neural network models are powerful empirical models approach, which can be compared with mathematical models. [source] APPLICATION OF GREY MODEL AND ARTIFICIAL NEURAL NETWORKS TO FLOOD FORECASTING,JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION, Issue 2 2006Moon Seong Rang ABSTRACT: The main focus of this study was to compare the Grey model and several artificial neural network (ANN) models for real time flood forecasting, including a comparison of the models for various lead times (ranging from one to six hours). For hydrological applications, the Grey model has the advantage that it can easily be used in forecasting without assuming that forecast storm events exhibit the same stochastic characteristics as the storm events themselves. The major advantage of an ANN in rainfall-runoff modeling is that there is no requirement for any prior assumptions regarding the processes involved. The Grey model and three ANN models were applied to a 2,509 km2 watershed in the Republic of Korea to compare the results for real time flood forecasting with from one to six hours of lead time. The fifth-order Grey model and the ANN models with the optimal network architectures, represented by ANN1004 (34 input nodes, 21 hidden nodes, and 1 output node), ANN1010 (40 input nodes, 25 hidden nodes, and 1 output node), and ANN1004T (14 input nodes, 21 hidden nodes, and 1 output node), were adopted to evaluate the effects of time lags and differences between area mean and point rainfall. The Grey model and the ANN models, which provided reliable forecasts with one to six hours of lead time, were calibrated and their datasets validated. The results showed that the Grey model and the ANN1010 model achieved the highest level of performance in forecasting runoff for one to six lead hours. The ANN model architectures (ANN1004 and ANN1010) that used point rainfall data performed better than the model that used mean rainfall data (ANN1004T) in the real time forecasting. The selected models thus appear to be a useful tool for flood forecasting in Korea. [source] PREDICTION OF LOCAL SCOUR AROUND BRIDGE PIERS USING ARTIFICIAL NEURAL NETWORKS,JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION, Issue 2 2006Sung-Uk Choi ABSTRACT: This paper describes a method for predicting local scour around bridge piers using an artificial neural network (ANN). Methods for selecting input variables, calibrations of network control parameters, learning process, and verifications are also discussed. The ANN model trained by laboratory data is applied to both laboratory and field measurements. The results illustrate that the ANN model can be used to predict local scour in the laboratories and in the field better than other empirical relationships that are currently in use. A parameter study is also carried out to investigate the importance of each input variable as reflected in data. [source] BACKLASH COMPENSATION IN NONLINEAR SYSTEMS USING DYNAMIC INVERSION BY NEURAL NETWORKSASIAN JOURNAL OF CONTROL, Issue 2 2000Rastko R. Selmic ABSTRACT A dynamic inversion compensation scheme is presented for backlash. The compensator uses the backstepping technique with neural networks (NN) for inverting the backlash nonlinearity in the feedforward path. The technique provides a general procedure for using NN to determine the dynamic preinverse of an invertible dynamical system. A tuning algorithm is presented for the NN backlash compensator which yields a stable closed-loop system. [source] Dynamic Wavelet Neural Network for Nonlinear Identification of Highrise BuildingsCOMPUTER-AIDED CIVIL AND INFRASTRUCTURE ENGINEERING, Issue 5 2005Xiaomo Jiang Compared with conventional neural networks, training of a dynamic neural network for system identification of large-scale structures is substantially more complicated and time consuming because both input and output of the network are not single valued but involve thousands of time steps. In this article, an adaptive Levenberg,Marquardt least-squares algorithm with a backtracking inexact linear search scheme is presented for training of the dynamic fuzzy WNN model. The approach avoids the second-order differentiation required in the Gauss,Newton algorithm and overcomes the numerical instabilities encountered in the steepest descent algorithm with improved learning convergence rate and high computational efficiency. The model is applied to two highrise moment-resisting building structures, taking into account their geometric nonlinearities. Validation results demonstrate that the new methodology provides an efficient and accurate tool for nonlinear system identification of high-rising buildings. [source] Position-Invariant Neural Network for Digital Pavement Crack AnalysisCOMPUTER-AIDED CIVIL AND INFRASTRUCTURE ENGINEERING, Issue 2 2004Byoung Jik Lee This system includes three neural networks: (1) image-based neural network, (2) histogram-based neural network, and (3) proximity-based neural network. These three neural networks were developed to classify various crack types based on the subimages (crack tiles) rather than crack pixels in digital pavement images. These spatial neural networks were trained using artificially generated data following the Federal Highway Administration (FHWA) guidelines. The optimal architecture of each neural network was determined based on the testing results from different sets of the number of hidden units, learning coefficients, and the number of training epochs. To validate the system, actual pavement pictures taken from pavements as well as the computer-generated data were used. The proximity value is determined by computing relative distribution of crack tiles within the image. The proximity-based neural network effectively searches the patterns of various crack types in both horizontal and vertical directions while maintaining its position invariance. The final result indicates that the proximity-based neural network produced the best result with the accuracy of 95.2% despite its simplest neural network structure with the least computing requirement. [source] Nonparametric Identification of a Building Structure from Experimental Data Using Wavelet Neural NetworkCOMPUTER-AIDED CIVIL AND INFRASTRUCTURE ENGINEERING, Issue 5 2003Shih-Lin Hung By combining wavelet decomposition and artificial neural networks (ANN), wavelet neural networks (WNN) are used for solving chaotic signal processing. The basic operations and training method of wavelet neural networks are briefly introduced, since these networks can approximate universal functions. The feasibility of structural behavior modeling and the possibility of structural health monitoring using wavelet neural networks are investigated. The practical application of a wavelet neural network to the structural dynamic modeling of a building frame in shaking tests is considered in an example. Structural acceleration responses under various levels of the strength of the Kobe earthquake were used to train and then test the WNNs. The results reveal that the WNNs not only identify the structural dynamic model, but also can be applied to monitor the health condition of a building structure under strong external excitation. [source] Probabilistic Neural Network for Reliability Assessment of Oil and Gas PipelinesCOMPUTER-AIDED CIVIL AND INFRASTRUCTURE ENGINEERING, Issue 5 2002Sunil K. Sinha A fuzzy artificial neural network (ANN),based approach is proposed for reliability assessment of oil and gas pipelines. The proposed ANN model is trained with field observation data collected using magnetic flux leakage (MFL) tools to characterize the actual condition of aging pipelines vulnerable to metal loss corrosion. The objective of this paper is to develop a simulation-based probabilistic neural network model to estimate the probability of failure of aging pipelines vulnerable to corrosion. The approach is to transform a simulation-based probabilistic analysis framework to estimate the pipeline reliability into an adaptable connectionist representation, using supervised training to initialize the weights so that the adaptable neural network predicts the probability of failure for oil and gas pipelines. This ANN model uses eight pipe parameters as input variables. The output variable is the probability of failure. The proposed method is generic, and it can be applied to several decision problems related with the maintenance of aging engineering systems. [source] An Adaptive Conjugate Gradient Neural Network,Wavelet Model for Traffic Incident DetectionCOMPUTER-AIDED CIVIL AND INFRASTRUCTURE ENGINEERING, Issue 4 2000H. Adeli Artificial neural networks are known to be effective in solving problems involving pattern recognition and classification. The traffic incident-detection problem can be viewed as recognizing incident patterns from incident-free patterns. A neural network classifier has to be trained first using incident and incident-free traffic data. The dimensionality of the training input data is high, and the embedded incident characteristics are not easily detectable. In this article we present a computational model for automatic traffic incident detection using discrete wavelet transform, linear discriminant analysis, and neural networks. Wavelet transform and linear discriminant analysis are used for feature extraction, denoising, and effective preprocessing of data before an adaptive neural network model is used to make the traffic incident detection. Simulated as well as actual traffic data are used to test the model. For incidents with a duration of more than 5 minutes, the incident-detection model yields a detection rate of nearly 100 percent and a false-alarm rate of about 1 percent for two- or three-lane freeways. [source] Tensammetric Analysis of Nonionic Surfactant Mixtures by Artificial Neural NetworkELECTROANALYSIS, Issue 12 2005A. Safavi Abstract An artificial neural network (ANN) model has been developed for tensammetric determination of a series of Brijes (Brij 30, Brij 35, Brij 56, Brij 96) as nonionic surfactants. The tensammetric method is based on the measurement of the capacitive current of the mercury electrode after adsorption of surfactants. All Brijes were analyzed in the concentration range of 1.0,100.0,,g mL,1. The proposed method shows good sensitivity and applicability to the simultaneous determination of mixtures of four Brijes in aqueous solutions. [source] Action control of autonomous agents in continuous valued space using RFCNELECTRONICS & COMMUNICATIONS IN JAPAN, Issue 2 2008Shinichi Shirakawa Abstract Researchers on action control of autonomous agents and multiple agents have attracted increasing attention in recent years. The general methods using action control of agents are neural network, genetic programming, and reinforcement learning. In this study, we use neural network for action control of autonomous agents. Our method determines the structure and parameter of neural network in evolution. We proposed Flexibly Connected Neural Network (FCN) previously as a method of constructing arbitrary neural networks with optimized structures and parameters to solve unknown problems. FCN was applied to action control of an autonomous agent and showed experimentally that it is effective for perceptual aliasing problems. All of the experiments of FCN, however, are only in grid space. In this paper, we propose a new method based on FCN which can decide correction action in real and continuous valued space. The proposed method, called Real-valued FCN (RFCN), optimizes input,output functions of each unit, parameters of the input,output functions and speed of each unit. In order to examine its effectiveness, we applied the proposed method to action control of an autonomous agent to solve continuous-valued maze problems. © 2008 Wiley Periodicals, Inc. Electron Comm Jpn, 91(2): 31,39, 2008; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/eej.10032 [source] Automated Evaluation of Kikuchi Patterns by Means of Radon and Fast Fourier Transformation, and Verification by an Artificial Neural Network,ADVANCED ENGINEERING MATERIALS, Issue 8 2003R.A. Schwarzer Automated crystal orientation measurement (ACOM) in the SEM by interpreting backscatter Kikuchi patterns (see Figure) has become a standard tool of quantitative texture analysis in materials science during the last decade. A Radon transformation of the diffraction pattern, in combination with a 1D fast Fourier transformation, enables the fast extraction of the positions of Kikuchi bands. The high-frequency coefficients of the 1D FFT are used to define pattern quality as a measure of lattice imperfection and residual stress of the real crystal structure. [source] Evaluation of in silico splice tools for decision-making in molecular diagnosis,HUMAN MUTATION, Issue 7 2008Claude Houdayer Abstract It appears that all types of genomic nucleotide variations can be deleterious by affecting normal pre-mRNA splicing via disruption/creation of splice site consensus sequences. As it is neither pertinent nor realistic to perform functional testing for all of these variants, it is important to identify those that could lead to a splice defect in order to restrict transcript analyses to the most appropriate cases. Web-based tools designed to provide such predictions are available. We evaluated the performance of six of these tools (Splice Site Prediction by Neural Network [NNSplice], Splice-Site Finder [SSF], MaxEntScan [MES], Automated Splice-Site Analyses [ASSA], Exonic Splicing Enhancer [ESE] Finder, and Relative Enhancer and Silencer Classification by Unanimous Enrichment [RESCUE]-ESE) using 39 unrelated retinoblastoma patients carrying different RB1 variants (31 intronic and eight exonic). These 39 patients were screened for abnormal splicing using puromycin-treated cell lines and the results were compared to the predictions. As expected, 17 variants impacting canonical AG/GT splice sites were correctly predicted as deleterious. A total of 22 variations occurring at loosely defined positions (±60 nucleotides from an AG/GT site) led to a splice defect in 19 cases and 16 of them were classified as deleterious by at least one tool (84% sensitivity). In other words, three variants escaped in silico detection and the remaining three were correctly predicted as neutral. Overall our results suggest that a combination of complementary in silico tools is necessary to guide molecular geneticists (balance between the time and cost required by RNA analysis and the risk of missing a deleterious mutation) because the weaknesses of one in silico tool may be overcome by the results of another tool. Hum Mutat 29(7), 975,982, 2008. © 2008 Wiley-Liss, Inc. [source] Applications of Sinusoidal Neural Network and Momentum Genetic Algorithm to Two-wheel Vehicle Regulating ProblemIEEJ TRANSACTIONS ON ELECTRICAL AND ELECTRONIC ENGINEERING, Issue 1 2008Duong Chau Sam Non-member Abstract In an attempt to enhance the performance of neural network (NN), we propose a sinusoidal activation function for NN and apply a fast genetic algorithm (GA) with uses of momentum offspring (MOS) and constant-range mutation (CRM) for training the NN. The proposed methods are aimed at designing a neurocontroller (NC) for regulating a two-wheel vehicle system, known as nonholonomic system, in the viewpoint that it is necessary to improve the control process of the system even though several control methods, including applications of NN and GAs, have been developed. The learning performances of NCs are evaluated through the successful evolutionary rates of the control process based on the values of the squared errors. In order to compare the conventional methods with our proposed approaches and verify the effects of momentum GA on NC training, various numerical simulations will be carried out with different numbers of generations in GAs and different activation functions of NCs. Finally, the controllability of NC is investigated with certain sets of initial states. The simulations show that sinusoidal NC trained by momentum GA has a good performance regardless of the small values of population size and generations in GA. Copyright © 2007 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc. [source] Generalized strain probing of constitutive modelsINTERNATIONAL JOURNAL FOR NUMERICAL AND ANALYTICAL METHODS IN GEOMECHANICS, Issue 15 2004Youssef M. A. Hashash Abstract Advanced material constitutive models are used to describe complex soil behaviour. These models are often used in the solution of boundary value problems under general loading conditions. Users and developers of constitutive models need to methodically investigate the represented soil response under a wide range of loading conditions. This paper presents a systematic procedure for probing constitutive models. A general incremental strain probe, 6D hyperspherical strain probe (HSP), is introduced to examine rate-independent model response under all possible strain loading conditions. Two special cases of HSP, the true triaxial strain probe (TTSP) and the plane-strain strain probe (PSSP), are used to generate 3-D objects that represent model stress response to probing. The TTSP, PSSP and general HSP procedures are demonstrated using elasto-plastic models. The objects resulting from the probing procedure readily highlight important model characteristics including anisotropy, yielding, hardening, softening and failure. The PSSP procedure is applied to a Neural Network (NN) based constitutive model. It shows that this probing is especially useful in understanding NN constitutive models, which do not contain explicit functions for yield surface, hardening, or anisotropy. Copyright © 2004 John Wiley & Sons, Ltd. [source] Use of neural networks for the prediction of frictional drag and transmission of axial load in horizontal wellboresINTERNATIONAL JOURNAL FOR NUMERICAL AND ANALYTICAL METHODS IN GEOMECHANICS, Issue 2 2003Tanvir Sadiq Abstract The use of mud motors and other tools to accomplish forward motion of the bit in extended reach and horizontal wells allows avoiding large amounts of torque caused by rotation of the whole drill string. The forward motion of the drill string, however, is resisted by excessive amount of friction. In the presence of large compressive axial loads, the drill pipe or coiled tubing tends to buckle into a helix in horizontal boreholes. This causes additional frictional drag resisting the transmission of axial load (resulting from surface slack-off force) to the bit. As the magnitude of the frictional drag increases, a buckled pipe may become ,locked-up' making it almost impossible to drill further. In case of packers, the frictional drag may inhibit the transmission of set-up load to the packer. A prior knowledge of the magnitude of frictional drag for a given axial load and radial clearance can help avoid lock-up conditions and costly failure of the tubular. In this study a neural network model, for the prediction of frictional drag and axial load transmission in horizontal wellbores, is presented. Several neural network architectures were designed and tested to obtain the most accurate prediction. After cross-validation of the Back Propagation Neural Network (BPNN) algorithm, a two-hidden layer model was chosen for simultaneous prediction of frictional drag and axial load transmission. A comparison of results obtained from BPNN and General Regression Neural Network (GRNN) algorithms is also presented. Copyright © 2002 John Wiley & Sons, Ltd. [source] Adaptive recurrent neural network control of biological wastewater treatmentINTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS, Issue 2 2005Ieroham S. Baruch Three adaptive neural network control structures to regulate a biological wastewater treatment process are introduced: indirect, inverse model, and direct adaptive neural control. The objective is to keep the concentration of the recycled biomass proportional to the influent flow rate in the presence of periodically acting disturbances, process parameter variations, and measurement noise. This is achieved by the so-called Jordan Canonical Recurrent Trainable Neural Network, which is a completely parallel and parametric neural structure, permitting the use of the obtained parameters, during the learning phase, directly for control system design. Comparative simulation results confirmed the applicability of the proposed control schemes. © 2005 Wiley Periodicals, Inc. Int J Int Syst 20: 173,193, 2005. [source] Multiobjective Tabu Search method used in chemistryINTERNATIONAL JOURNAL OF QUANTUM CHEMISTRY, Issue 6 2006T. Rusu Abstract The use of a combined artificial intelligence method in macromolecular chemistry design is described. This method implies a Back-Propagation (BP) Neural Network, modified for two-dimensional input data and for a system composed of a genetic algorithm extended by a Tabu Search operator used to incorporate high-level chemical knowledge: thermodynamic polymer properties. © 2005 Wiley Periodicals, Inc. Int J Quantum Chem, 2006 [source] Assessment of Individual Risk of Death Using Self-Report Data: An Artificial Neural Network Compared with a Frailty IndexJOURNAL OF AMERICAN GERIATRICS SOCIETY, Issue 7 2004Xiaowei Song PhD Objectives: To evaluate the potential of an artificial neural network (ANN) in predicting survival in elderly Canadians, using self-report data. Design: Cohort study with up to 72 months follow-up. Setting: Forty self-reported characteristics were obtained from the community sample of the Canadian Study of Health and Aging. An individual frailty index score was calculated as the proportion of deficits experienced. For the ANN, randomly selected participants formed the training sample to derive relationships between the variables and survival and the validation sample to control overfitting. An ANN output was generated for each subject. A separate testing sample was used to evaluate the accuracy of prediction. Participants: A total of 8,547 Canadians aged 65 to 99, of whom 1,865 died during 72 months of follow-up. Measurements: The output of an ANN model was compared with an unweighted frailty index in predicting survival patterns using receiver operating characteristic (ROC) curves. Results: The area under the ROC curve was 86% for the ANN and 62% for the frailty index. At the optimal ROC value, the accuracy of the frailty index was 70.0%. The ANN accuracy rate over 10 simulations in predicting the probability of individual survival mean±standard deviation was 79.2±0.8%. Conclusion: An ANN provided more accurate survival classification than an unweighted frailty index. The data suggest that the concept of biological redundancy might be operationalized from health survey data. [source] Sensory Modeling of Coffee with a Fuzzy Neural NetworkJOURNAL OF FOOD SCIENCE, Issue 1 2002O. Tominaga ABSTRACT: Models were constructed to predict sensory evaluation scores from the blending ratio of coffee beans. Twenty-two blended coffees were prepared from 3 representative beans and were evaluated with respect to 10 sensory attributes by 5 coffee cup-tasters and by models constructed using the response surface method (RSM), multiple regression analysis (MRA), and a fuzzy neural network (FNN). The RSM and MRA models showed good correlations for some sensory attributes, but lacked sufficient overall accuracy. The FNN model exhibited high correlations for all attributes, clearly demonstrated the relationships between blending ratio and flavor characteristics, and was accurate enough for practical use. FNN, thus, constitutes a powerful tool for accelerating product development. [source] A Comparison of Neural Network, Statistical Methods, and Variable Choice for Life Insurers' Financial Distress PredictionJOURNAL OF RISK AND INSURANCE, Issue 3 2006Patrick L. Brockett This study examines the effect of the statistical/mathematical model selected and the variable set considered on the ability to identify financially troubled life insurers. Models considered are two artificial neural network methods (back-propagation and learning vector quantization (LVQ)) and two more standard statistical methods (multiple discriminant analysis and logistic regression analysis). The variable sets considered are the insurance regulatory information system (IRIS) variables, the financial analysis solvency tracking (FAST) variables, and Texas early warning information system (EWIS) variables, and a data set consisting of twenty-two variables selected by us in conjunction with the research staff at TDI and a review of the insolvency prediction literature. The results show that the back-propagation (BP) and LVQ outperform the traditional statistical approaches for all four variable sets with a consistent superiority across the two different evaluation criteria (total misclassification cost and resubstitution risk criteria), and that the twenty-two variables and the Texas EWIS variable sets are more efficient than the IRIS and the FAST variable sets for identification of financially troubled life insurers in most comparisons. [source] Application of an Artificial Neural Network for Simulating Robust Plasma-Sprayed Zirconia CoatingsJOURNAL OF THE AMERICAN CERAMIC SOCIETY, Issue 5 2008Ming-Der Jean This article presents the application of the artificial neural network (ANN) of a statistically designed experiment for developing a robust wear-resistant zirconia coating. In this research, experimental design with orthogonal arrays efficiently provides enough information with the least number of experiments, reducing the cost and time. A radial basis function (RBF) network for the wear behavior is adopted. The friction and tribological properties of zirconia coatings were investigated. The microstructural feature of the coatings is also addressed in this study. It is found that the worn volumes of plasma-sprayed zirconia coatings after wear tests are greatly improved by the optimal parameters. The relationships between the microstructure of the worn surface and their properties are examined, and the results reveal a higher wear resistance and a lower worn surface roughness with a large amount of plastic deformations. These wear resistant structures formed as a result of a dense lamellar formation during sprayed zirconia coatings. The RBF network can be established efficiently. A comparison of the predicted results with that of the RBF network and the Taguchi method predictor shows average errors of 2.735% and 9.191% for the RBF network and the Taguchi method, respectively. It is experimentally confirmed that the RBF network predictions are in agreement with the experiments, and it can be reliably used for the prediction of wear for plasma sprayings. The experimental results demonstrate that the RBF network used for a statistically designed experiment is an effective, efficient, and intelligent approach for developing a robust, high efficiency, and high-quality zirconia coating process. [source] FLOOD STAGE FORECASTING WITH SUPPORT VECTOR MACHINES,JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION, Issue 1 2002Shie-Yui Liong ABSTRACT: Machine learning techniques are finding more and more applications in the field of forecasting. A novel regression technique, called Support Vector Machine (SVM), based on the statistical learning theory is explored in this study. SVM is based on the principle of Structural Risk Minimization as opposed to the principle of Empirical Risk Minimization espoused by conventional regression techniques. The flood data at Dhaka, Bangladesh, are used in this study to demonstrate the forecasting capabilities of SVM. The result is compared with that of Artificial Neural Network (ANN) based model for one-lead day to seven-lead day forecasting. The improvements in maximum predicted water level errors by SVM over ANN for four-lead day to seven-lead day are 9.6 cm, 22.6 cm, 4.9 cm and 15.7 cm, respectively. The result shows that the prediction accuracy of SVM is at least as good as and in some cases (particularly at higher lead days) actually better than that of ANN, yet it offers advantages over many of the limitations of ANN, for example in arriving at ANN's optimal network architecture and choosing useful training set. Thus, SVM appears to be a very promising prediction tool. [source] A neural network-based approach to determine FDTD eigenfunctions in quantum devicesMICROWAVE AND OPTICAL TECHNOLOGY LETTERS, Issue 9 2009Antonio Soriano Abstract This article combines a Neural Network (NN) algorithm with the Finite Difference Time Domain (FDTD) technique to estimate the eigenfunctions in quantum devices. A NN based on the Least Mean Squares (LMS) algorithm is combined with the FDTD technique to provide a first approach to the confined states in quantum wires. The proposed technique is in good agreement with analytical results and is more efficient than FDTD combined with the Fourier Transform. This technique is used to calculate a numerical approximation to the eigenfunctions associated to quantum wire potentials. The performance and convergence of the proposed technique are also presented in this article. © 2009 Wiley Periodicals, Inc. Microwave Opt Technol Lett 51: 2017,2022, 2009; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop.24562 [source] Characterization of Mixtures Part 1: Prediction of Infinite-Dilution Activity Coefficients Using Neural Network-Based QSPR ModelsMOLECULAR INFORMATICS, Issue 11-12 2008Subhash Ajmani Abstract The major problem in building QSAR/QSPR models for mixtures lies in their characterization. It has been shown that it is possible to construct QSPR models for the density of binary liquid mixtures using simple mole fraction weighted physicochemical descriptors. Such parameters are unsatisfactory; however, from the point of view of interpretation of the resultant models. In this paper, an alternative mechanism-based approach to the characterization of mixtures has been investigated. It has been shown that while it is not possible to build significant linear models using these descriptors, it has been possible to construct satisfactory artificial neural network models. The performance of these models and the importance of the individual descriptors are discussed. [source] Dynamic Process Modelling using a PCA-based Output Integrated Recurrent Neural NetworkTHE CANADIAN JOURNAL OF CHEMICAL ENGINEERING, Issue 4 2002Yu Qian Abstract A new methodology for modelling of dynamic process systems, the output integrated recurrent neural network (OIRNN), is presented in this paper. OIRNN can be regarded as a modified Jordan recurrent neural network, in which the past values for certain steps of the output variables are integrated with the input variables, and the original input variables are pre-processed using principal component analysis (PCA) for the purpose of dimension reduction. The main advantage of the PCA-based OIRNN is that the input dimension is reduced, so that the network can be used to model the dynamic behavior of multiple input multiple output (MIMO) systems effectively. The new method is illustrated with reference to the Tennessee-Eastman process and compared with principal component regression and feedforward neural networks. On présente dans cet article une nouvelle méthodologie pour la modélisation de systèmes de procédés dynamiques, soit le réseau neuronal récurrent avec intégration de la réponse (OIRNN). Ce dernier peut être vu comme un réseau neuronal récurrent de Jordan modifié, dans lequel les valeurs passées pour certaines étapes des valeurs de sortie sont intégrées aux variables d'entrée et les variables d'entrée originales pré-traitée par l'analyse des composants principaux (PCA) dans un but de réduction des dimensions. Le principal avantage de l'OIRNN basé sur la PCA est que la dimension d'entée est réduite de sorte que le réseau peut servir à modéliser le comportement dynamique de systèmes à entrée et sorties multiples (MIMO) de façon efficace. La nouvell méthod est illustrée dans le cas du procédé Tennessee-Eastman et est comparée aux réseaux neuronaux anticipés et à régression des composants principaux. [source] | ||||||||||||||||||||||||||||||||||||||||||||||||||||