Multilayer Backpropagation Neural Network Research Articles

Earthquake performance of infilled frames using neural networks and experimental database

Reinforced-concrete frames with masonry wall infill, “framed-masonry”, is a composite structural system proven to be effective and efficient in the case of in plane horizontal excitations. Its behaviour depends on mechanical characteristics of its components but its performance is different than the sum of its components. Modelling and seismic design verifications of “framed-masonry” system that embraces all of the important aspects of behaviour, failure mechanism, shear strength and deformation capacity, are required. In this work we have tried to put the “frame–masonry” composite as a full-fledged building element whose performance could be determined quantitatively on the basis of data obtained from the performed tests. Frame–masonry composite was analyzed using neural networks trained on the experimental database that contains results of 113 published tests of one-story one-bay masonry infilled frames. In order to reduce the dimensionality of input data and achieve better performance of neural network, dimensionality reduction techniques: Principal Component Analysis, Forward stepwise sensitivity analysis and dimensionless modelling parameter approach were applied. A multilayered back propagation neural network with adaptive weight function was applied and the optimal network topology, for each required output value, was been chosen. The obtained results indicated that neural network, trained on the database, could be used for predicting the seismic behaviour of framed-masonry structural elements, with limitation of inputs according to the statistical range of input data. Sensitivity analysis of the important factors that affect the performance indicated that the most important ones were height/length ratio (a), material properties of masonry infill and frame (fk, fck), reinforcement ratio of columns (rc) and the amount of vertical loading (N).

Optimization of the Composite Truncated Cone Structure Layers under Buckling Load

The orientation of fibers in the layers is an important factor that must be obtained in order to predict how well the finished composite product will perform under real-world working conditions. In this research, a five-layer glass-epoxy composite truncated cone structure under buckling load was considered. The simulation of the structure was done utilizing finite element method and was confirmed comparing with the published experimental results. Then the effect of different orientation of fibers on the buckling load was considered. For this, a computer programing was developed to compute the buckling load for different orientations of fibers in each layer. These orientations were produced randomly with the delicacy of 15 degrees. Finally, neural network and genetic algorithm methods were utilized to obtain the optimum orientations of fibers in each layer using the training data obtained from finite element simulation. There are many parameters such as the number of hidden layers, the number of neurons in each hidden layer, the training algorithm, the activation function and so on which must be specified properly in development of a neural network model. The number of hidden layers and number of neurons in each layer was obtained by try and error method. In this study, multilayer back-propagation (BP) neural network with the Levenberg-Marquardt training algorithm (trainlm) was used. Finally, the results showed that the truncated cone with optimum layers withstand considerably more buckling load.

The worst response of mistuned bladed disk system using neural network and genetic algorithm

The objective of this paper is to develop an integrated approach using artificial neural networks (ANN) and genetic algorithms (GA) for predicting the worst response of mistuned bladed disk. ANN is used to predict the responses of bladed disk system which are used further in evaluation of fitness and constraint violation in GA process. A multilayer back-propagation neural network is trained with the results obtained from finite element model for different bladed disk configurations. Subsequently, GA is employed for arriving at optimum configuration of the bladed disk system by maximizing the blade responses. By integrating ANN with GA, the computational time required for obtaining optimal solution could be reduced substantially. The efficacy of this approach is demonstrated by carrying out studies on mistuned bladed disk systems for different sets of mistuning parameters, namely mistuning in modulus of elasticity and length of blades. Finally, the effect of adding shroud at the tip of blades in reducing the maximum response of the bladed disk system was investigated.

Carotid artery image segmentation using modified spatial fuzzy c-means and ensemble clustering

Disease diagnosis based on ultrasound imaging is popular because of its non-invasive nature. However, ultrasound imaging system produces low quality images due to the presence of spackle noise and wave interferences. This shortcoming requires a considerable effort from experts to diagnose a disease from the carotid artery ultrasound images. Image segmentation is one of the techniques, which can help efficiently in diagnosing a disease from the carotid artery ultrasound images. Most of the pixels in an image are highly correlated. Considering the spatial information of surrounding pixels in the process of image segmentation may further improve the results. When data is highly correlated, one pixel may belong to more than one clusters with different degree of membership. In this paper, we present an image segmentation technique namely improved spatial fuzzy c-means and an ensemble clustering approach for carotid artery ultrasound images to identify the presence of plaque. Spatial, wavelets and gray level co-occurrence matrix (GLCM) features are extracted from carotid artery ultrasound images. Redundant and less important features are removed from the features set using genetic search process. Finally, segmentation process is performed on optimal or reduced features. Ensemble clustering with reduced feature set outperforms with respect to segmentation time as well as clustering accuracy. Intima-media thickness (IMT) is measured from the images segmented by the proposed approach. Based on IMT measured values, Multi-Layer Back-Propagation Neural Networks (MLBPNN) is used to classify the images into normal or abnormal. Experimental results show the learning capability of MLBPNN classifier and validate the effectiveness of our proposed technique. The proposed approach of segmentation and classification of carotid artery ultrasound images seems to be very useful for detection of plaque in carotid artery.

Contrast-enhanced ultrasonography parameters in neural network diagnosis of liver tumors

To study the role of time-intensity curve (TIC) analysis parameters in a complex system of neural networks designed to classify liver tumors. We prospectively included 112 patients with hepatocellular carcinoma (HCC) (n = 41), hypervascular (n = 20) and hypovascular (n = 12) liver metastases, hepatic hemangiomas (n = 16) or focal fatty changes (n = 23) who underwent contrast-enhanced ultrasonography in the Research Center of Gastroenterology and Hepatology, Craiova, Romania. We recorded full length movies of all contrast uptake phases and post-processed them offline by selecting two areas of interest (one for the tumor and one for the healthy surrounding parenchyma) and consecutive TIC analysis. The difference in maximum intensities, the time to reaching them and the aspect of the late/portal phase, as quantified by the neural network and a ratio between median intensities of the central and peripheral areas were analyzed by a feed forward back propagation multi-layer neural network which was trained to classify data into five distinct classes, corresponding to each type of liver lesion. The neural network had 94.45% training accuracy (95% CI: 89.31%-97.21%) and 87.12% testing accuracy (95% CI: 86.83%-93.17%). The automatic classification process registered 93.2% sensitivity, 89.7% specificity, 94.42% positive predictive value and 87.57% negative predictive value. The artificial neural networks (ANN) incorrectly classified as hemangyomas three HCC cases and two hypervascular metastases, while in turn misclassifying four liver hemangyomas as HCC (one case) and hypervascular metastases (three cases). Comparatively, human interpretation of TICs showed 94.1% sensitivity, 90.7% specificity, 95.11% positive predictive value and 88.89% negative predictive value. The accuracy and specificity of the ANN diagnosis system was similar to that of human interpretation of the TICs (P = 0.225 and P = 0.451, respectively). Hepatocellular carcinoma cases showed contrast uptake during the arterial phase followed by wash-out in the portal and first seconds of the late phases. For the hypovascular metastases did not show significant contrast uptake during the arterial phase, which resulted in negative differences between the maximum intensities. We registered wash-out in the late phase for most of the hypervascular metastases. Liver hemangiomas had contrast uptake in the arterial phase without agent wash-out in the portal-late phases. The focal fatty changes did not show any differences from surrounding liver parenchyma, resulting in similar TIC patterns and extracted parameters. Neural network analysis of contrast-enhanced ultrasonography - obtained TICs seems a promising field of development for future techniques, providing fast and reliable diagnostic aid for the clinician.

An assessment of electrical load forecasting using artificial neural network

The forecasting of electricity demand has become one of the major research fields in electrical engineering. The supply industry requires forecasts with lead times, which range from the short term (a few minutes, hours, or days ahead) to the long term (up to 20 years ahead). The major priority for an electrical power utility is to provide uninterrupted power supply to its customers. Long term peak load forecasting plays an important role in electrical power systems in terms of policy planning and budget allocation. This paper presents a peak load forecasting model using artificial neural networks (ANN). The approach in the paper is based on multi-layered back-propagation feed forward neural network. For annual forecasts, there should be 10 to 12 years of historical monthly data available for each electrical system or electrical buss. A case study is performed by using the proposed method of peak load data of a state electricity board of India which maintain high quality, reliable, historical data providing the best possible results. Model's quality is directly dependent upon data integrity.

Shannon Entropy and Mean Square Errors for speeding the convergence of Multilayer Neural Networks: A comparative approach

Improving the efficiency and convergence rate of the Multilayer Backpropagation Neural Network Algorithms is an active area of research. The last years have witnessed an increasing attention to entropy based criteria in adaptive systems. Several principles were proposed based on the maximization or minimization of entropic cost functions. One way of entropy criteria in learning systems is to minimize the entropy of the error between two variables: typically one is the output of the learning system and the other is the target. In this paper, improving the efficiency and convergence rate of Multilayer Backpropagation (BP) Neural Networks was proposed. The usual Mean Square Error (MSE) minimization principle is substituted by the minimization of Shannon Entropy (SE) of the differences between the multilayer perceptions output and the desired target. These two cost functions are studied, analyzed and tested with two different activation functions namely, the Cauchy and the hyperbolic tangent activation functions. The comparative approach indicates that the Degree of convergence using Shannon Entropy cost function is higher than its counterpart using MSE and that MSE speeds the convergence than Shannon Entropy.

Crude oil price forecasting: Experimental evidence from wavelet decomposition and neural network modeling

Oil price prediction has usually proved to be an intractable task due to the intrinsic complexity of oil market mechanism. In addition, the recent oil shock and its consequences relaunch the debate on understanding the behavior underlying the expected oil prices. Combining the dynamic properties of multilayer back propagation neural network and the recent Harr A trous wavelet decomposition, a Hybrid model HTW-MPNN is implemented to achieve prominent prediction of crude oil price. While recent studies focus on the determination of the best forecasting model by comparing various neural architectures or applying several decomposition techniques to the ANN, the new insight of this paper is to target the issue of the transfer function selection providing robust simulations on both in sample and out of sample basis. Based on the work of Yonaba, H., Anctil, F., and Fortin, V. (2010) “Comparing Sigmoid Transfer Functions for Neural Network Multistep Ahead Stream flow forecasting”. Journal of Hydrologic Engineering, April, 275–283, we use three variants of activation function namely sigmoid, bipolar sigmoid and hyperbolic tangent in order to test the model's flexibility. Furthermore, the forecasting robustness is checked through several levels of input–hidden nodes. Comparatively, results of HTW-MBPNN perform better than the conventional BPNN. Our conclusions add a major attribute to the previous studies corroborating the Occam razor's principle, especially when simulations are constructed through training and testing phases simultaneously. Finally, more eligible forecasting power is found according to the wavelet oil price signal which appears to be the closest to the real anticipations of future oil price fluctuations.

Neural Network to Identify Individuals at Health Risk

The risk of diseases such as heart attack and high blood pressure could be reduced by adequate physical activity. However, even though majority of general population claims to perform some physical exercise, only a minority exercises enough to keep a healthy living style. Thus, physical inactivity has become one of the major concerns of public health in the past decade. Research shows that the highest decrease in physical activity is noticed from high school to college. Thus, it is of great importance to quickly identify college students at health risk due to physical inactivity. Research also shows that the level of physical activity of an individual is highly correlated to demographic features such as race and gender, as well as self motivation and support from family and friends. This information could be collected from each student via a 20 minute questionnaire, but the time needed to distribute and analyze each questionnaire is infeasible on a collegiate campus. Thus, we propose an automatic identifier of students at risk, so that these students could easier be targeted by collegiate campuses and physical activity promotion departments. We present in this paper preliminary results of a supervised backpropagation multilayer neural network for classifying students into at-risk or not at-risk group .

Phase equilibria prediction of solid solute in supercritical carbon dioxide with and without a cosolvent: The use of artificial neural network

In this study, a back-propagation multi-layer neural network was developed to predict the solubility of solid solute in supercritical carbon dioxide with and without cosolvent. The solubility of anthracene in CO 2 with cosolvents, acetone, ethanol and cyclohexane were employed as model systems to investigate the supercritical carbon dioxide behaviour in ternary systems over a wide range of temperatures. The back-propagation neural network operated in a supervised learning mode. A number of networks were trained and tested with different network parameters using training and testing data sets. To establish the network applicability, a validating data set was used and the predictability of the network was statistically evaluated. Statistical estimations showed that the neural network predictions had an excellent agreement with experimental data. The calculated average relative deviation (ARD) and the root mean squared error (RMSD) for tested ANNs data points were 5.45% and 0.74%, respectively. A minimum number of data points have been employed to train the ANN. The predicted ARD and RMSD for the employed ternary systems were 7.83% and 0.07%, respectively. The results obtained in this work indicate that ANN is a superior technique with high level of accuracy for prediction of solubility of solid solute in ternary systems.

Knowledge-based evolving connectionist systems for condition evaluation of sustainable roadways: a feasibility study

In place, sustainable rehabilitation of existing deteriorated concrete highways through rubblization is considered to be a green and economical alternative to other options involving total reconstruction, etc. There is currently no standardised method of backcalculating the rubblized pavement layer moduli from pavement non-destructive test data through inverse analysis. This paper explores the feasibility of applying two evolving connectionist systems (ECOS), namely dynamic evolving neuro-fuzzy inference system (DENFIS) and evolving fuzzy neural network (EFuNN), to rubblized pavement moduli backcalculation. It is advantageous to employ ECOS networks for analysing complex engineered systems such as rubblized pavements since ECOS networks are resistant to catastrophic forgetting, have the ability to adapt to and learn new data as soon as they become available, do not have a limit to the amount of knowledge they can store and learn the examples very quickly compared to traditional multi-layered perceptron backpropagation neural networks (MLP-BP NN).

Analysis of international debt problem using artificial neural networks and statistical methods

It is known from the scientific researches that artificial neural networks are alternatives of statistical methods such as regression analysis and classification in recent years. Since multi-layer backpropagation neural network models are nonlinear, it is expected that the neural network models should make better classifications and predictions. The studies on this subject support that idea. In this study, a macro-economic problem on rescheduling or non-rescheduling of the countries’ international debts is taken into account. Among the statistical methods, logistic and probit regression, and the different neural network backpropagation algorithms are applied and comparisons are made. Evaluations and suggestions are made depending on the results and different neural network architecture.

A Comparison of Time Series Forecasting using Support Vector Machine and Artificial Neural Network Model

Time series prediction is an important problem in many applications in natural science, engineering and economics. The objective of this study is to examine the flexibility of Support Vector Machine (SVM) in time series forecasting by comparing it with a multi-layer back-propagation (BP) neural network. Five well-known time series data sets are used in this study to demonstrate the effectiveness of the forecasting model. These data are utilized to forecast through an application aimed to handle real life time series. The grid search technique using 10-fold cross validation is used to determine the best value of SVM parameters in the forecasting process. The experiment shows that SVM outperforms the BP neural network based on the criteria of Mean Absolute Error (MAE). It also indicates that SVM provides a promising technique in time series forecasting techniques.

An Improved Medical Decision Support System to Identify the Breast Cancer Using Mammogram

An improved Computer Aided Clinical Decision Support System has been developed to classify the tumor and identify the stages of the cancer using neural network and presented in this paper. The texture and shape features have been extracted and the optimal feature set has been obtained using multiobjective genetic algorithm (MOGA). The multilayer back propagation neural network with Ant Colony Optimization and Particle Swarm Optimization has been used. The accuracy of the proposed system has been verified and found that the accuracy of 99.5% can be achieved. The proposed system can provide valuable information to the physicians in clinical pathology.

Back Propagation Neural Network-Modified Covariance Modelling for Improvement of Holographic Imaging

Two approaches are combined to improve the performance of modelling long wavelength holographic imaging (detection) of concealed objects. One approach is to design a multi-layer back propagation neural network that is able to reduce the effect of noise in a captured signal and results in a model as close as possible to the desired one. The other approach is to further process the captured signal by applying a modified covariance spectral estimation method to improve the resolution of the reconstructed image. Different concealing media and different values of signal to noise ratio are used to investigate the performance of such combination. It has been proved that the proposed model can provide an overall improvement in the imaging of concealed objects.

Performance comparison of three artificial neural network methods for classification of electroencephalograph signals of five mental tasks

In this paper, performance of three classifiers for classification of five mental tasks were investigated. Wavelet Packet Transform (WPT) was used for feature extraction of the relevant frequency bands from raw Electroencephalograph (EEG) signal. The three classifiers namely used were Multilayer Back propagation Neural Network, Support Vector Machine and Radial Basis Function Neural Network. In MLP-BP NN five training methods used were a) Gradient Descent Back Propagation b) Levenberg-Marquardt c) Resilient Back Propagation d) Conjugate Learning Gradient Back Propagation and e) Gradient Descent Back Propagation with movementum.

An integrated approach for optimum design of HPC mix proportion using genetic algorithm and artificial neural networks

This study aims to develop a cost-based high-performance concrete (HPC) mix optimization system based on an integrated approach using artificial neural networks (ANNs) and genetic algorithms (GA). ANNs are used to predict the three main properties of HPC, namely workability, strength and durability, which are used to evaluate fitness and constraint violations in the GA process. Multilayer back-propagation neural networks are trained using the results obtained from experiments and previous research. The correlation between concrete components and its properties is established. GA is employed to arrive at an optimal mix proportion of HPC by minimizing its total cost. A system prototype, called High Performance Concrete Mix-Design System using Genetic Algorithm and Neural Networks (HPCGANN), was developed in MATLAB. The architecture of the proposed system consists of three main parts: 1) User interface; 2) ANNs prediction models software; and 3) GA engine software. The validation of the proposed system is carried out by comparing the results obtained from the system with the trial batches. The results indicate that the proposed system can be used to enable the design of HPC mix which corresponds to its required performance. Furthermore, the proposed system takes into account the influence of the fluctuating unit price of materials in order to achieve the lowest cost of concrete, which cannot be easily obtained by traditional methods or trial-and-error techniques.

A multistage knowledge-based system for EEG seizure detection in newborn infants

Objective Automatic seizure detection has attracted attention as a method to obtain valuable information concerning the duration, timing, and frequency of seizures. Methods currently used to detect EEG seizures in adults show high false detection rates in neonates because they lack information about specific age-dependent features of normal and pathological EEG and artifacts. This paper describes a novel multistage knowledge-based seizure detection system for newborn infants to identify and classify normal and pathological newborn EEGs as well as seizures with a reduced false detection rate. Methods We developed the system in a way to make comprehensive use of spatial and temporal contextual information obtained from multichannel EEGs. The system development consists of six major stages: (i) EEG data collection and bandpass filtering; (ii) automatic artifact detection; (iii) feature extraction from segments of non-seizure and seizure activities; (iv) feature selection via the relevance and redundancy analysis; (v) EEG classification and pattern recognition using a trained multilayer back-propagation neural network; and (v) knowledge-based decision-making to examine each of possible EEG patterns from a multi-channel perspective. The system was developed and tested with the EEG recordings of 10 newborns aged between 39 and 42 weeks. Results The overall sensitivity, selectivity, and average detection rate of the system were 74%, 70.1%, and 79.7%, respectively. The average false detection of 1.55/h was also achieved by the system with a feature reduction up to 80%. Conclusions The expert rule-based decision-making subsystem accompanying the classifier helped to reduce the false detection rate, reject a wide variety of artifacts, and discriminate various patterns of EEG. Significance This paper may serve as a guide for the selection of discriminative features to improve the accuracy of conventional seizure detection systems for routine clinical EEG interpretation and brain activity monitoring in newborns especially those hospitalized in the neonatal intensive care units.

An integrated approach for optimum design of bridge decks using genetic algorithms and artificial neural networks

The objective of this paper is to develop an integrated approach using artificial neural networks (ANN) and genetic algorithms (GA) for cost optimization of bridge deck configurations. In the present work, ANN is used to predict the structural design responses which are used further in evaluation of fitness and constraint violation in GA process. A multilayer back-propagation neural network is trained with the results obtained using grillage analysis program for different bridge deck configurations and the correlation between sectional parameters and design responses has been established. Subsequently, GA is employed for arriving at optimum configuration of the bridge deck system by minimizing the total cost. By integrating ANN with GA, the computational time required for obtaining optimal solution could be reduced substantially. The efficacy of this approach is demonstrated by carrying out studies on cost optimization of T-girder bridge deck system for different spans. The method presented in this paper, would greatly reduce the computational effort required to find the optimum solution and guarantees bridge engineers to arrive at the near-optimal solution that could not be easily obtained using general modeling programs or by trial-and-error.

Coronary Artery Diagnosis Aided by Neural Network

Coronary Artery Diagnosis Aided by Neural Network Coronary artery disease is due to atheromatous narrowing and subsequent occlusion of the coronary vessel. Application of optimised feed forward multi-layer back propagation neural network (MLBP) for detection of narrowing in coronary artery vessels is presented in this paper. The research was performed using 580 data records from traditional ECG exercise test confirmed by coronary arteriography results. Each record of training database included description of the state of a patient providing input data for the neural network. Level and slope of ST segment of a 12 lead ECG signal recorded at rest and after effort (48 floating point values) was the main component of input data for neural network was. Coronary arteriography results (verified the existence or absence of more than 50% stenosis of the particular coronary vessels) were used as a correct neural network training output pattern. More than 96% of cases were correctly recognised by especially optimised and a thoroughly verified neural network. Leave one out method was used for neural network verification so 580 data records could be used for training as well as for verification of neural network.