Contents of Volume 25 (2015)

1/2015 2/2015 3/2015 4/2015 5/2015


  • [1] Editorial, 585-586,    Full text

  • [2] Yolcu U., Bas E., Egrioglu E., Aladag C.H. (Turkey)
    A New Multilayer Feedforward Network Based on Trimmed Mean Neuron Model, 587-602

    First page   Full text     DOI: 10.14311/NNW.2015.25.029

    Abstract: The multilayer perceptron model has been suggested as an alternative to conventional approaches, and can accurately forecast time series. Additionally, several novel artificial neural network models have been proposed as alternatives to the multilayer perceptron model, which have used (for example) the generalized- mean, geometric mean, and multiplicative neuron models. Although all of these artificial neural network models can produce successful forecasts, their aggregation functions mean that they are negatively affected by outliers. In this study, we propose a new multilayer, feed forward neural network model, which is a robust model that uses the trimmed mean neuron model. Its aggregation function does not depend on outliers. We trained this multilayer, feed forward neural network using modied particle swarm optimization. We applied the proposed method to three well-known time series, and our results suggest that it produces superior forecasts when compared with similar methods.

  • [3] Cakit E., Durgun B., Cetik O.(Turkey)
    A Neural Network Approach for Assessing the Relationship between Grip Strength and Hand Anthropometry, 603-622

    First page   Full text     DOI: 10.14311/NNW.2015.25.030

    Abstract: This study aimed to determine grip strength data for Turkish dentistry students and developed prediction models that allow: i) investigation of the relationship between grip strength and hand anthropometry using artificial neural networks (ANNs) and stepwise regression analysis, ii) prediction of the grip strength of Turkish dentistry students, and iii) assessment of the potential impact of hand anthropometric variables on grip strength. The study included 153 right-handed dentistry students, consisting of 81 males and 72 females. From 44 anthropometric and biomechanical measurements obtained from the right hands of the participants; five anthropometric measurements were selected for ANN and regression modeling using stepwise regression analysis. We included stepwise regression analysis results to assess the predictive power of the neural network approach, in comparison to a classical statistical approach. When the model accuracy was calculated based on the coefficient of determination (R2), the root mean squared error (RMSE) and the mean absolute error (MAE) values for each of the models, ANN showed greater predictive accuracy than regression analysis, as demonstrated by experimental results. For the best performing ANN model, the testing values of the models correlated well with actual values, with a coefficient of determination (R2) of 0.858. Using the best performing ANN model, sensitivity analysis was applied to determine the effects of hand dimensions on grip strength and to rank these dimensions in order of importance. The results suggest that the three most sensitive input variables are the forearm length, the hand breadth and the finger circumference at the first joint of digit 5 and that the ANNs are promising techniques for predicting hand grip strength based on hand breadth, finger breadth, hand length, finger circumference and forearm length.

  • [4] Wang N., Ma Y., Wang W., Zhan K.(China)
    Multifocus Image Fusion Based on Nonsubsampled Contourlet Transform and Spiking Cortical Model, 623-639

    First page   Full text     DOI: 10.14311/NNW.2015.25.031

    Abstract: A novel image fusion algorithm based on nonsubsampled contourlet transform (NSCT) and spiking cortical model (SCM) is proposed in this paper, aiming at solving the fusion problem of multifocus images. The fusion rules of subband coefficients of NSCT are discussed, and a new maximum selection rule (MSR) is defined to fuse low frequency coefficients instead of using traditional MSR directly. For the fusion rule of high frequency coefficients, spatial frequency (SF) of each high frequency subband is considered as the gradient features of images to motivate SCM networks and generate pulse of neurons, and then the time matrix of SCM is set as criteria to select coefficients of high frequency subband. Experimental results and visual evaluation demonstrate the effectiveness of the proposed fusion method. Objective tests and analysis conducted under different noised source image environments proved the robustness of the proposed fusion method.

  • [5] Wang J., Liu S., Wang H., Zeng Y. (China)
    Dynamical Properties of Firing Patterns in the Huber-Braun Cold Receptor Model in Response to External Current Stimuli, 641-655

    First page   Full text     DOI: 10.14311/NNW.2015.25.032

    Abstract: We have studied the role of external current stimuli in a four-dimensional Hodgkin-Huxley-type model of cold receptor in this paper. Firstly, we researched its firing patterns from direct current (DC) and alternating current (AC) stimuli. Under different values of DC stimulus intensity, interspike intervals (ISIs) with period-doubling bifurcation phenomena appeared. Second, research has shown that neurons are extremely sensitive to changes in the frequency and amplitude of the current used to stimulate them. As the stimulus frequency increased, discharge rhythms emerged ranging from burst firing to chaotic firing and spiking firing. Meanwhile, various phase-locking patterns have been studied in this paper, such as p : 1 (p > 1), 1 : q (q > 1), 2 : q (q > 1) and p : q (p; q > 1), etc. Finally, based on the fast-slow dynamics analysis, codimension-two bifurcation analysis of the fast subsystem was performed in the parameter (asr;B)-plane. We mainly investigated cusp bifurcation, fold-Hopf bifurcation, Bogdanov-Takens bifurcation and generalized Hopf bifurcation. These results revealed the effect of external current stimuli on the neuronal discharge rhythm and were instructive for further understanding the dynamical properties and mechanisms of the Huber-Braun model.

  • [6] Zhang Y., Cai G., Sun J., Wang Y., Chen J. (China)
    A New Sparse Low-rank Matrix Decomposition Method and its Application on Train Passenger Abnormal Action Identification, 657-668

    First page   Full text     DOI: 10.14311/NNW.2015.25.033

    Abstract: In the article a new sparse low-rank matrix decomposition model is proposed based on the smoothly clipped absolute deviation (SCAD) penalty. In order to overcome the computational hurdle we generalize the alternating direction method of multipliers (ADMM) algorithm to develop an alternative algorithm to solve the model. The algorithm we designed alternatively renew the sparse matrix and low-rank matrix in terms of the closed form of SCAD penalty. Thus, the algorithm reduces the computational complexity while at the same time to keep the computational accuracy. A series of simulations have been designed to demonstrate the performances of the algorithm with comparing with the Augmented Lagrange Multiplier (ALM) algorithm. Ultimately, we apply the model to an on- board video background modeling problem. According to model the on-board video background, we can separate the video background and passenger's actions. Thus, the model can help us to identify the abnormal action of train passengers. The experiments show the background matrix we estimated is not only sparser, but the computational efficiency is also improved.

  • [7] Tian Z., Li S., Wang Y., Wang X. (China)
    A Network Traffic Hybrid Prediction Model Optimized by Improved Harmony Search Algorithm, 669-685

    First page   Full text     DOI: 10.14311/NNW.2015.25.034

    Abstract: The telecommunication and Ethernet trafic prediction problem is studied. Network traffic prediction is an important problem of telecommunication and Ethernet congestion control and network management. In order to improve network traffic prediction accuracy, a network traffic hybrid prediction model was proposed by using the advantages of grey model and Elman neural network, grey model and Elman neural network predictive values were independently obtained, the different weight coefficients of two prediction models were given. In terms of weight coefficients optimization, an improved harmony search algorithm with better convergence speed and accuracy was proposed, the optimal weight coefficients of network traffic hybrid prediction model were determined through this algorithm, two prediction models results were multiplied by the weight coefficients to obtain the final prediction value. The network traffic sample data from an actual telecommunication network was collected as simulation object. The simulation results verified that the proposed network traffic hybrid prediction model based on improved harmony search algorithm has higher prediction accuracy.

  • [8] Contents volume 25 (2015), ... 687
  • [9] Authors index volume 25 (2015), ...691


  • [1] Basterrech S., Rubino G. (CZ, FR)
    Random Neural Network Model for Supervised Learning Problems, 457-500

    First page   Full text     DOI: 10.14311/NNW.2015.25.024

    Abstract: Random Neural Networks (RNNs) area classof Neural Networks (NNs) that can also be seen as a specific type of queuing network. They have been successfully used in several domains during the last 25 years, as queuing networks to analyze the performance of resource sharing in many engineering areas, as learning tools and in combinatorial optimization, where they are seen as neural systems, and also as models of neurological aspects of living beings. In this article we focus on their learning capabilities, and more specifically, we present a practical guide for using the RNN to solve supervised learning problems. We give a general description of these models using almost indistinctly the terminology of Queuing Theory and the neural one. We present the standard learning procedures usedby RNNs, adapted from similar well-established improvements in the standard NN field. We describe in particular a set of learning algorithms covering techniques based on the use of first order and, then, of second order derivatives. We also discuss some issues related to these objects and present new perspectives about their use in supervised learning problems. The tutorial describes their most relevant applications, and also provides a large bibliography.

  • [2] Simsek H., Cemek B., Odabas M.S., Rahman S. (Turkey)
    Estimation of Nutrient Concentrations in Runoff from Beef Cattle Feedlot using Adaptive Neuro-Fuzzy Inference Systems, 501-518

    First page   Full text     DOI: 10.14311/NNW.2015.25.025

    Abstract: Nutrient concentrations in runoff from beef cattle feedlots were estimated using two different adaptive network-based fuzzy inference systems (ANFIS), which were: (1) grid partition (ANFIS-GP) and (2) subtractive clustering based fuzzy inference system (ANFIS-SC). The input parameters were pH and electrical conductivity (EC); and the output parameters were total Kjeldahl nitrogen (TKN), ammonium-N (NH4-N), orthophosphate (ortho-P), and potassium (K). Models performances were evaluated based on root mean square error, mean absolute error, mean bias error, and determination coeficient statistics. For the same dataset, the ANFIS model outputs were also compared with a previously published nutrient concentration predictability model for runoff using artificial neural network (ANN) outputs. Results showed that both ANFIS-GP and ANFIS-SC models successfully predicted the runoff nutrient concentration. The comparison results revealed that the ANFIS-GP model performed slightly better than ANFIS-SC model in estimating TKN, NH4-N, ortho-P, and K. When compared with the ANN model for the same dataset, ANFIS outperformed ANN in nutrient concentration prediction in runoff.

  • [3] Chen Z., Zhu Y., Di Y., Feng S., Geng J. (China)
    A High-accuracy Self-adaptive Resource Demands Predicting Method in IaaS Cloud Environment, 519-540

    First page   Full text     DOI: 10.14311/NNW.2015.25.026

    Abstract: In IaaS (Infrastructure as a Service) cloud environment, users are provisioned with virtual machines (VMs). However, the initialization and resource allocation of virtual machines are not instantaneous and usually minutes of time are needed. Therefore, to realize efficient resource provision, it is necessary to know the accurate amount of resources needed to be allocated in advance. For this purpose, this paper proposes a high-accuracy self-adaptive prediction method using optimized neural network. The characters of users’ demands and preferences are analyzed firstly. To deal with the specific circumstances, a dynamic self-adaptive prediction model is adopted. Some basic predictors are adopted for resource requirements prediction of simple circumstances. BP neural network with self-adjusting learning rate and momentum is adopted to optimize the prediction results. High-accuracy self-adaptive prediction is realized by using the prediction results of basic predictors with different weights as training data besides the historical data. Feedback control is introduced to improve the whole operation performance. Statistic validation of the method is conducted adopting multiple evaluation criteria. The experiment results show that the method is promising for effectively predicting resource requirements in the cloud environment.

  • [4] Muthukumar B., Sivatha Sindhu S.S., Geetha S., Kannan A. (India)
    Intelligent Network-Misuse-Detection-System Using Neurotree Classifier, 541-564

    First page   Full text     DOI: 10.14311/NNW.2015.25.027

    Abstract: Intrusion detection systems (IDSs) are designed to distinguish normal and intrusive activities. A critical part of the IDS design depends on the selection of informative features and the appropriate machine learning technique. In this paper, we investigated the problem of IDS from these two perspectives and constructed a misuse based neurotree classiffier capable of detecting anomalies in networks. The major implications of this paper are a) Employing weighted sum genetic feature extraction process which provides better discrimination ability for detecting anomalies in network trafic; b) Realizing the system as a rule-based model using an ensemble efficient machine learning technique, neurotree which possesses better comprehensibility and generalization ability; c) Utilizing an activation function which is targeted at minimizing the error rates in the learning algorithm. An extensive experimental evaluation on a database containing normal and anomaly trafic patterns shows that the proposed scheme with the selected features and the chosen classiffier is a state-of-the-art IDS that outperforms previous IDS methods.

  • [5] Chiroma H., Abdul-kareem S., Ibrahim U., Gadam Ahmad I., Garba A., Abubakar A., Fatihu Hamza M., Herawan T. (Malaysia, Nigeria)
    Malaria Severity Classiffication through Jordan-Elman Neural Network Based on Features Extracted from Thick Blood Smear, 565-583

    First page   Full text     DOI: 10.14311/NNW.2015.25.028

    Abstract: This article presents an alternative approach useful for medical practitioners who wish to detect malaria and accurately identify the level of severity. Malaria classi?ers are usually based on feed forward neural networks. In this study, the proposed classiffier is developed based on the Jordan-Elman neural networks. Its performance is evaluated using a receiver-operating characteristic curve, sensitivity, specificity, positive predictive value, negative predictive value, confusion matrix, mean square error, determinant coefficient, and reliability. The effectiveness of the classiffier is compared to a support vector machine and multiple regression models. The results of the comparative analysis demonstrate a superior performance level of the Jordan-Elman neural network model. Further comparison of the classier with previous literature indicates performance improvement over existing results. The Jordan-Elman neural networks classiffier can assist medical practitioners in the fast detection of malaria and determining its severity, especially in tropical and subtropical regions where cases of malaria are prevalent


  • [1] M. Vaitová, P. Štemberk (CZ)
    Estimation of ablation depth in concrete slab under reactor during nuclear accident, 347-368

    First page   Full text     DOI: 10.14311/NNW.2015.25.018

    Abstract: The molten reactor core-concrete interaction, which describes the effect of molten reactor spread on the concrete oor of the reactor cavity, is a very complex process to simulate and predict, but the knowledge of this process is of major importance for planning the emergency counteractions for severe accidents with respect to the Stress Tests requirements after the Fukushima-Daiichi accident. The key issue is to predict the rate and most probable focusation of the melt-through process which is affected by the concrete composition, especially by the aggregate type. A limited number of small-scale experiments have been conducted over the past years along with accompanying numerical models which focused mainly on the siliceous type of aggregate. It is common for the concrete structures that the limestone type or the mixture of these two types of aggregate are used as well. Then, the objective of this paper is to extend the knowledge gained from the experiments with the siliceous aggregate to the concrete structures which are made of limestone aggregate or their combination, such as limestone sand and siliceous gravel. The proposed one-dimensional model of the melt-through process is based on the fuzzy-logic interpretation of the thermodynamic trends which reflect the aggregate type. This approach allows estimating the asymptotic cases in terms of the melt-through depth in the concrete oor over time with respect to the aggregate type, which may help to decide the rather expensive further experimental efforts.

  • [2] Krömer P., Prílepok M., Platoš J., Snášel V. (CZ)
    Graph visualisation by concurrent differential evolution , 369-386

    First page   Full text     DOI: 10.14311/NNW.2015.25.019

    Abstract: A representative dimensionality reduction is an important step in the analysis of real-world data. Vast amounts of raw data are generated by cyberphysical and information systems in different domains. They often feature a combination of high dimensionality, large volume, and vague, loosely defined structure. The main goal of visual data analysis is an intuitive, comprehensible, efficient, and graphically appealing representation of information and knowledge that can be found in such collections. In order to achieve an efficient visualisation, raw data need to be transformed into a refined form suitable for machine and human analysis. Various methods of dimension reduction and projection to low-dimensional spaces are used to accomplish this task. Sammon's projection is a well-known non-linear projection algorithm valued for its ability to preserve dependencies from an original high-dimensional data space in the low-dimensional projection space. Recently, it has been shown that bio-inspired real-parameter optimization methods can be used to implement the Sammon's projection on data from the domain of social networks. This work investigates the ability of several advanced types of the differential evolution algorithm as well as their parallel variants to minimize the error function of the Sammon's projection and compares their results and performance to a traditional heuristic algorithm.

  • [3] Ozsoydan F.B., Kandemir C.M., Demirtas E.A. (Turkey)
    Neural-network-based genetic algorithm for optimal kitchen faucet styles, 387-404

    First page   Full text     DOI: 10.14311/NNW.2015.25.020

    Abstract: Artificial neural networks (ANNs) are the models of choice in many data classification tasks. In this study, ANN classification models were used to explore user perceptions about kitchen faucet styles and investigate the relations between the overall preferences and kansei word scores of users. The scores given by consumers were obtained via a two-stage questionnaire mentioned in a previous study by the authors. Through the questionnaire, consumers were asked to give scores after examining three-dimensional (3-D) drawings of new product samples created with the help of industrial product designers. Because it was neither practical nor necessary to develop a prototype or a picture of each of the alternative designs, a fractional factorial experimental design similar to Taguchi's L-16 orthogonal array was used. After completing this preparatory work to develop ANNs and obtain the necessary related data, an analysis of variance (ANOVA) was performed to identify the critical factors that affect the accuracy of the ANN model to be used and determine the best factor levels for the ANN model. A genetic algorithm (GA) was then integrated with the ANN model found to be the best and implemented to determine the optimal levels of the design parameters related to product appearance. Lastly, the product categories were classified as unfavorable or favorable, and three products were derived for each category. In comparison with the previously published papers of the authors, the GA integrated with the ANN model was found to be an effective tool for revealing user perceptions in new product development. In regard to the findings of the present work, it can be said that, this technique can be used as an alternative of several complex analytical approaches, in order to explore users' perceptions.

  • [4] Alkama S., Chahir Y., Berkani D. (Algeria)
    Label maps fusion for the marginal segmentation of multi-component images, 405-426

    First page   Full text     DOI: 10.14311/NNW.2015.25.021

    Abstract: In this paper, we propose a new technique for merging the label maps obtained by the marginal segmentation of a multi-component image. In the marginal segmentation, each component of the multi-component image is independently segmented by labeling the pixels of the same class with the same label. Therefore the number of label maps corresponds to the number of components in the image. It is then necessary to merge them in order to have a single label map, i.e. a single segmented image. In the most merging techniques, the compatibility links between these maps are performed a priori by making the correspondences between their labels. However the various components are segmented and labeled independently, label maps are considered as independent sources. It is then difficult to establish the relationship compatibilities between labels. The method we propose does not a priori assume any compatibility links. The label maps are combined by superposition. Unfortunately, an over-segmentation is produced. To cope with this problem, the insignificant regions and classes are eliminated. Finally, classes are grouped by using hierarchical agglomerative clustering algorithm. Tests performed on color and satellite images show the effectiveness of this method and its superiority compared to the vector segmentation. The self-organizing map is used during the segmentation process in both marginal and vector segmentations.

  • [5] Rahmani A., Ghanbari A., Mahboubkhah M. (Iran)
    Kinematics analysis and numerical simulation of hybrid serial-parallel manipulator based on neural network, 427-442

    First page   Full text     DOI: 10.14311/NNW.2015.25.022

    Abstract: This paper presents solution of kinematics analysis of a specific class of serial-parallel manipulators, known as 2-(6UPS) manipulators, which are composed of several modules consisting of elementary manipulators with the parallel structure of the Stewart platform, by artificial neural network. At first, the kinematics model of the hybrid manipulator is obtained. Then, as the inverse kinematics problem of this kind of manipulators is a very difficult problem to solve because of their highly nonlinear relations between joint variables and position and orientation of the end effectors, a wavelet based neural network (wave-net) with its inherent learning ability as a strong method was used to solve the inverse kinematics problem. Also, the proposed wavelet neural network (WNN) is applied to approximate the paths of middle and upper plate in a circular and a spiral path, respectively. The results show high accurate performance of the proposed WNN.

  • [6] Tanidir O., Tör O.B. (Turkey)
    Accuracy of ANN based day-ahead load forecasting in Turkish power system: degrading and improving factors , 443-456

    First page   Full text     DOI: 10.14311/NNW.2015.25.023

    Abstract: This paper presents development of a day ahead load forecasting (DALF) model for Turkish power system with an artificial neural network (ANN). Effects of special holidays including national and religious days, and hourly random load deviations observed in Turkish power system due to significant arc furnace loads are discussed. Performance of the ANN is investigated in the sense of both DALF performance - in terms of both daily mean absolute percentage error (MAPE) and hourly absolute percentage error (APE) - and hourly secondary reserves required to ensure supply/demand adequacy of the system. The most sensitive cities to DALF in terms of daily city temperature forecasts are ranked in order to reduce the input of the developed ANN and thereby to improve execution of the model. Candidate cities are determined based on both their placement with respect to climatic zones of the country and their contribution to the system load during peak hours. The results show that, although a well-trained ANN could provide very satisfactory daily MAPEs at non-special days, such as ~1%, the hourly absolute percentage errors (APE) could be significant due to large random load disturbances, which necessitate special attention during the day ahead allocation of hourly secondary reserves. By limiting the temperature data set with major cities, the input of ANN reduces significantly while not disturbing the MAPEs. Main contributions of the study are; addressing both benefits of the prioritizing the cities in a power system in the sense of their temperature forecast effects on the DALF performance and assessing the performance of DALF in the sense of necessary amount of secondary reserves in power systems which include significant random load deviations (e.g., large arc furnace loads).


  • [1] Swietlicka A., Gugala K., Jurkowlaniec A., Sniatala P., Rybarczyk A. (Poland)
    The stochastic, Markovian, Hodgkin-Huxley type of mathematical model of the neuron, 219-239

    First page   Full text     DOI: 10.14311/NNW.2015.25.012

    Abstract: The aim of this paper is to show how the Hodgkin-Huxley model of the neuron's membrane potential can be extended to a stochastic one. This extension can be done either by adding fluctuations to the equations of the model or by using Markov kinetic schemes' formalism. We are presenting a new extension of the model. This modification simplifies computational complexity of the neuron model especially when considering a hardware implementation. The hardware implemen- tation of the extended model as a system on a chip using a field-programmable gate array (FPGA) is demonstrated in this paper. The results confirm the reliability of the extended model presented here.

  • [2] Akay M.F., Aci C.I, Abut F. (Turkey)
    Predicting the performance measures of a 2-dimensional message passing multiprocessor architecture by using machine learning methods, 241-265

    First page   Full text     DOI: 10.14311/NNW.2015.25.013

    Abstract: 2-dimensional Simultaneous Optical Multiprocessor Exchange Bus (2D SOME-Bus) is a reliable, robust implementation of petaflops-performance computer architecture. In this paper, we develop models to predict the performance measures (i.e. average channel utilization, average channel waiting time, average network latency, average processor utilization and average input waiting time) of a message passing architecture interconnected by the 2D SOME-Bus by using Multi- layer Feed-forward Artificial Neural Network (MFANN), Support Vector Regression (SVR) and Multiple Linear Regression (MLR). OPNET Modeler is used to simulate the message passing 2D SOME-Bus multiprocessor architecture and to create the training and testing datasets. Using 10-fold cross validation, the performance of the prediction models have been evaluated using several performance metrics. The results show that the SVR model using the radial basis function kernel (SVR-RBF) yields the lowest prediction error among all models.

  • [3] Golasowski M., Martinovic J., Litschmannová M., Kuchař S., Podhorányi M. (CZ)
    Uncertainty modelling in Rainfall-Runoff simulations based on parallel Monte Carlo method, 267-286

    First page   Full text     DOI: 10.14311/NNW.2015.25.014

    Abstract: This article describes statistical evaluation of the computational model for precipitation forecast and proposes a method for uncertainty modelling of rainfall-runoff models in the Floreon+ system based on this evaluation. The Monte-Carlo simulation method is used for estimating possible river discharge and provides several confidence intervals that can support the decisions in operational disaster management. Experiments with other parameters of the model and their influence on final river discharge are also discussed.

  • [4] Fengwei Li (China)
    Isolated rupture degree of trees and gear graphs, 287-300

    First page   Full text     DOI: 10.14311/NNW.2015.25.015

    Abstract: The isolated rupture degree for a connected graph G is defined as ir(G) = max{i(G-S)-|S|-m(G-S):S is element C(G)}, where i(G-S) and m(G-S), respectively, denote the number of components which are isolated vertices and the order of a largest component in G-S. C(G) denotes the set of all cut-sets of G. The isolated rupture degree is a new graph parameter which can be used to measure the vulnerability of networks. In this paper, we firstly give a recursive algorithm for computing the isolated rupture degree of trees, and determine the maximum and minimum isolated rupture degree of trees with given order and maximum degree. Then, the exact value of isolated rupture degree of gear graphs are given. In the final, we determine the rupture degree of the Cartesian product of two special graphs and a special permutation graph.

  • [5] Dogan G., Arslan M.H., Ceylan M. (Turkey)
    Statistical feature extraction based on an ANN approach for estimating the compressive strength of concrete, 301-318

    First page   Full text     DOI: 10.14311/NNW.2015.25.016

    Abstract: Applications of artificial intelligence in engineering disciplines have become widespread and have provided alternative solutions to engineering problems. Image processing technology (IPT) and artificial neural networks (ANNs) are types of artificial intelligence methods. However, IPT and ANN have been used together in extremely few studies. In this study, these two methods were used to deter- mine the compressive strength of concrete, a complex material whose mechanical features are difficult to predict. Sixty cube-shaped specimens were manufactured, and images of specific features of the specimens were taken before they were tested to determine their compressive strengths. An ANN model was constituted as a result of the process of digitizing the images. In this way, the two different artificial intelligence methods were used together to carry out the analysis. The compressive strength values of the concrete obtained via analytical modeling were compared with the test results. The results of the comparison (R² = 0:9837-0:9961) indicate that the combination of these two artificial intelligence methods is highly capable of predicting the compressive strengths of the specimens. The model's predictive capability was also evaluated in terms of several statistical parameters using a set of statistical methods during the digitization of the images constituting the artificial neural network.

  • [6] Sönmez F., Bülbül S. (Turkey)
    An intelligent software model design for estimating deposit banks profitability with soft computing techniques, 319-345

    First page   Full text     DOI: 10.14311/NNW.2015.25.017

    Abstract: Profitability of Turkish banking sector gained importance after national and international financial crisis happened in the last decade, which revealed the need to make a research on profitability and the factors determining profitability. In recent years, new techniques of soft computing (SC) like genetic algorithms (GAs), fuzzy logic (FL) and especially artificial neural networks (ANNs) have been applied into the financial domain to solve the domain issues because of their successful applications in nonlinear multivariate situations. An adaptive system was needed due to the fact that insufficient use of application software programs for SC and the fact that single software is only applicable for specific model. Furthermore, even though ANNs have been applied to many areas; little attention has been paid to estimation of bank profitability with ANNs. This article is intended to analyze and estimate the profitability of deposit banks in Turkey with an adaptive software model of ANNs which have not been previously applied for this context, comprehensively. The results from the software model, which processes the factors affecting profitability, indicate that all of the variables used have significant impacts in varying proportions on profitability and that obtained estimations achieved the targeted and acceptable performance of success. This software model is expected to provide easiness on estimating bank profitability, since giving successful estimations and not being affected by user differences. Additionally, it is aimed to construct a software model for being used in different fields of study and financial domain.


  • [1] Hyun Jun Park, Kwang Baek Kim, Eui Young Cha (South Korea)
    An Effective Color Quantization Method Using Color Importance-Based Self-Organizing Maps, 121-137

    First page   Full text     DOI: 10.14311/NNW.2015.25.006

    Abstract: Color quantization is an important process for image processing and various applications. Up to now, many color quantization methods have been proposed. The self-organizing maps (SOM) method is one of the most effective color quantization methods, which gives excellent color quantization results. However, it is slow, so it is not suitable for real-time applications. In this paper, we present a color importance{based SOM color quantization method. The proposed method dynamically adjusts the learning rate and the radius of the neighborhood using color importance. This makes the proposed method faster than the conventional SOM-based color quantization method. We compare the proposed method to 10 well-known color quantization methods to evaluate performance. The methods are compared by measuring mean absolute error (MAE), mean square error (MSE), and processing time. The experimental results show that the proposed method is effective and excellent for color quantization. Not only does the proposed method provide the best results compared to the other methods, but it uses only 67.18% of the processing time of the conventional SOM method.

  • [2] Özen Yelbasi, Emin Germen (Turkey)
    A Self Organizing Map Based Approach for Congestion Avoidance in Autonomous IP Networks, 139-160

    First page   Full text     DOI: 10.14311/NNW.2015.25.007

    Abstract: This work presents a Self Organizing Map (SOM) based queue management approach against congestion in autonomous Internet Protocol (IP) networks. The new queue management approach is proposed with consideration to the pros and cons of two well-known queue management algorithms: Random Early Detection (RED) and Drop Tail (DT). At the beginning of this study, RED and DT are compared by observing their effects on two important indicators of congestion: end-to-end delay and delay variation. This comparison reveals that the performances of RED and DT vary according to the level of global congestion: under low congestion conditions, when packet losses caused by congestion are unlikely, DT outperforms RED; while under high congestion, RED is superior to DT. The SOM based approach takes into account the variations in the global congestion levels and makes decisions to optimise congestion avoidance. A centralized observation unit is designed for monitoring global congestion levels in autonomous IP networks. A traffic flow is generated between each router and the observation unit so as to follow the changes in the global congestion level. For this purpose, IP routers are specialized to send packets carrying queue length information to the observation unit. A SOM based decision mechanism is used by the observation unit, to make predictions on the future congestion behavior of the network and inform the routers. Routers use this information to update their congestion avoidance behavior, as their ability to update their RED parameters is enhanced by the congestion notifications sent by the observation unit. In this work, multiple simulations are undertaken in order to test the performance of the proposed SOM-based method. A considerable improvement is observed from the point of view of end-to-end delays and delay variations, by comparison with DT and RED as used in recent IP networks.

  • [3] Marek Běhálek, Martin Šurkovský, Ondřej Meca, Stanislav Böhm (CZ)
    Memory Optimized Pheromone Structures for Max-Min Ant System, 161-174

    First page   Full text     DOI: 10.14311/NNW.2015.25.008

    Abstract: Ant Colony Optimization is a meta-heuristic for solving hard combinatorial optimization problems. It is a constructive population based approach inspired by the social behavior of ants. In our research, we are focused on the parallel/distributed computing on massively parallel systems. More precisely, we want to adjust Max-Min Ant System (one of Ant Colony Optimization algorithms) for these systems. Traditionally, a matrix is used to store the pheromone information. If we want to solve large instances, this is a very memory consuming solution. In this paper, we propose a different approach. We do not use the matrix to store the pheromone information. Instead, ant trails that are normally incorporated into this matrix are stored during the computation and just some parts and only in time when they are really needed are assembled. Proposed solution was implemented in C++. The implemented solution was tested on large symmetric instances of Traveling Salesman Problem. In these experiments, we were able to compute results with a comparable quality and even faster than with the traditional approach while using only a portion of the original memory.

  • [4] Chao Shao, Chunhong Wan, Haitao Hu (China)
    Manifold Learning and Visualization Based on Dynamic Self-Organizing Map, 175-188

    First page   Full text     DOI: 10.14311/NNW.2015.25.009

    Abstract: For the data sampled from a low-dimensional nonlinear manifold embedded in a high-dimensional space, such as Swiss roll and S-curve, Self-Organizing Map (SOM) tends to get stuck in local minima and then yield topological defects in the final map. To avoid this problem and obtain more faithful visualization results, a variant of SOM, i.e. Dynamic Self-Organizing Map (DSOM), was presented in this paper. DSOM can dynamically increase the map size, as the training data set is expanded according to its intrinsic neighborhood structure, starting from a small neighborhood in which the data points can lie on or close to a linear patch. According to the locally Euclidean nature of the manifold, the map can be guided onto the manifold surface and then the global faithful visualization results can be achieved step by step. Experimental results show that DSOM can discover the intrinsic manifold structure of the data more faithfully than SOM. In addition, as a new manifold learning method, DSOM can obtain more concise visualization results and be less sensitive to the neighborhood size and the noise than typical manifold learning methods, such as Isometric Mapping (ISOMAP) and Locally Linear Embedding (LLE), which can also be verified by experimental results.

  • [5] Xinhua Xue, Yangpeng Li, Xingguo Yang, Xin Chen, Jian Xiang (China)
    Prediction of Slope Stability Based on GA-BP Hybrid Algorithm, 189-202

    First page   Full text     DOI: 10.14311/NNW.2015.25.010

    Abstract: Safety monitoring and stability analysis of high slopes are important for high dam construction in mountainous regions or precipitous gorges. Slope stability estimation is an engineering problem that involves several parameters. To address these problems, a hybrid model based on the combination of Genetic algorithm (GA) and Back-propagation Artificial Neural Network (BP-ANN) is proposed in this study to improve the forecasting performance. GA was employed in selecting the best BP-ANN parameters to enhance the forecasting accuracy. Several important parameters, including the slope geological conditions, location of instruments, space and time conditions before and after measuring, were used as the input parameters, while the slope displacement was the output parameter. The results shown that the GA-BP model is a powerful computational tool that can be used to predict the slope stability.

  • [6] Helena Bínová (CZ)
    Modified Method of Gravity Model Application for Transatlantic Air Transportation, 203-217

    First page   Full text     DOI: 10.14311/NNW.2015.25.011

    Abstract: Air transportation between Europe and the U.S. is becoming more and more significant. It can only hardly be replaced by other means of transportation, since its biggest advantages include speed and reliability. Air transportation forecasting is important for planning the development of airports and related infrastructure, and of course also for air carriers. Therefore, it is important to forecast the number of flights between selected airports in Europe and the U.S. and the number of transported persons. A gravity model is usually used for this forecasting. Determination of coefficients which significantly affect results of the formulas used in the gravity model is crucial. Coefficients are, as a rule, computed by an iterative algorithm implementing the gradient method. This technique has some limitations if the state space is inappropriate. Moreover, the exponent parameter in the formula is obviously fixed. We have chosen the new method of differential evolution to determine the gravity model coefficient. Differential evolution works with populations similarly to other evolution algorithms. It is suitable for solving complex numerical problems. The suggested methodology can be helpful for various airlines to forecast demand and plan new long-haul routes.


  • [1] Svítek M. (CZ)
    TUTORIAL - Towards Complex System Theory, 5-33

    First page   Full text     DOI: 10.14311/NNW.2015.25.001

    Abstract: This tutorial summarizes the new approach to complex system theory that comes basically from physical information analogies. The information components and gates are defined in a similar way as components in electrical or mechanical engineering. Such approach enables the creation of complex networks through their serial, parallel or feedback ordering. Taking into account wave probabilistic functions in analogy with quantum physics, we can enrich the system theory with features such as entanglement. It is shown that such approach can explain emergencies and self-organization properties of complex systems.

  • [2] Přikryl J., Kocijan J. (CZ, Slovenia)
    Modelling Occupancy-Queue Relation Using Gaussian Process, 35-52

    First page   Full text     DOI: 10.14311/NNW.2015.25.002

    Abstract: One of the key indicators of the quality of service for urban transportation control systems is the queue length. Even in unsaturated conditions, longer queues indicate longer travel delays and higher fuel consumption. With the exception of some expensive surveillance equipment, the queue length itself cannot be measured automatically, and manual measurement is both impractical and costly in a long term scenario. Hence, many mathematical models that express the queue length as a function of detector measurements are used in engineering practice, ranging from simple to elaborate ones. The method proposed in this paper makes use of detector time-occupancy, a complementary quantity to vehicle count, provided by most of the traffic detectors at no cost and disregarded by majority of existing approaches for various reasons. Our model is designed as a complement to existing methods. It is based on Gaussian-process model of the occupancy-queue relationship, it can handle data uncertainties, and it provides more information about the quality of the queue length prediction.

  • [3] Aizhu Zhang, Genyun Sun, Zhenjie Wang, Yanjuan Yao (China)
    A Hybrid Genetic Algorithm and Gravitational Search Algorithm for Global Optimization, 53-73

    First page   Full text     DOI: 10.14311/NNW.2015.25.003

    Abstract: The laws of gravity and mass interactions inspire the gravitational search algorithm (GSA), which finds optimal regions of complex search spaces through the interaction of individuals in a population of particles. Although GSA has proven effective in both science and engineering, it is still easy to suffer from premature convergence especially facing complex problems. In this paper, we proposed a new hybrid algorithm by integrating genetic algorithm (GA) and GSA (GA-GSA) to avoid premature convergence and to improve the search ability of GSA. In GA-GSA, crossover and mutation operators are introduced from GA to GSA for jumping out of the local optima. To demonstrate the search ability of the proposed GA-GSA, 23 complex benchmark test functions were employed, including unimodal and multimodal high-dimensional test functions as well as multimodal test functions with fixed dimensions. Wilcoxon signed-rank tests were also utilized to execute statistical analysis of the results obtained by PSO, GSA, and GA-GSA. Experimental results demonstrated that the proposed algorithm is both efficient and effective.

  • [4] Daassi-Gnaba H., Oussar Y. (France)
    External vs. Internal SVM-RFE: The SVM-RFE Method Revisited and Applied to Emotion Recognition, 75-91

    First page   Full text     DOI: 10.14311/NNW.2015.25.004

    Abstract: Support Vector Machines (SVM) are well known as a kernel based method mostly applied to classification. SVM-Recursive Feature Elimination (SVM- RFE) is a variable ranking and selection method dedicated to the design of SVM based classifiers. In this paper, we propose to revisit the SVM-RFE method. We study two implementations of this feature selection method that we call External SVM-RFE and Internal SVM-RFE, respectively. The two implementations are applied to rank and select acoustic features extracted from speech to design optimized linear SVM classifiers that recognize speaker emotions. To show the efficiency of the External and Internal SVM-RFE methods, an extensive experimental study is presented. The SVM classifiers were selected using a validation procedure that ensures strict speaker independence. The results are discussed and compared with those achieved when the features are ranked using the Gram-Schmidt procedure. Overall, the results achieve a recognition rate that exceeds 90%.

  • [5] Feng Guo, Lin Lin, Xiaolong Xie, Bin Luo (China)
    Novel Hybrid Rule Network Based on TS Fuzzy Rules, 93-116

    First page   Full text     DOI: 10.14311/NNW.2015.25.005

    Abstract: A novel hybrid rule network based on TS fuzzy rules is proposed to resolve the problems of fuzzy classification and prediction. The proposed model learns by using genetic algorithm and is able to cover the whole distribution regions of the samples. In the learning process: (1) fuzzy intervals of each dimension of the samples are partitioned evenly; (2) computing intervals (CIs) are established based on the even intervals; (3) linear weighted model of several normal probability distributions is used to describe the sample probability distribution on CIs; (4) membership degree of each CI is learnt to evaluate the importance of each CI, avoiding the problem that the optimal intervals are difficult to cover the original sample spaces; (5) dynamic rule selection mechanism is used to dynamically combine a small number of optimal rules linearly to achieve nonlinear approximation, reducing the computation load. Three experiments are performed: the experiments on Iris and Mackey-Glass chaotic time series show that HRN can achieve satisfactory results and is more effective in terms of generalization ability, whereas the experiment on exhaust gas temperature demonstrates that HRN can predict the EGT of aero engine effectively.