Contents of Volume 30 (2020)

1/2020 2/2020 3/2020 4/2020 5/2020

6/2020

  • [1] Mehralian S., Teshnehlab M., Nasersharif B., (Iran)
    Traffic data analysis using deep Elman and gated recurrent auto-encoder, pp. 347-363

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.023

    Abstract: Traffic flow prediction is one of the most interesting machine learning applications in real-world problems that can help anyone move around. In this study, we proposed a feature extraction structure for multivariate time series using Elman recurrent auto-encoder. We added loopback from the encoder layer of the normal auto-encoder to regard sequence information between successive data. The feedback layer implemented using Elman neural network and GRU cells, then the model is trained by diflerent optimization algorithms. The models are also trained using the Emotional Learning method in which we involve the derivative of the error in the cost function to avoid local minimums and keep the last state of the network. We used the proposed method for classification and prediction problems on traffic data from the California Department of Transportation Performance Measurement System (PeMS). The results show that our structure can successfully extract a compact representation of traffic data useful for reconstructing of original data, classification, and prediction. The results also show that adding the recurrent layer to the feature extractor (auto-encoder) leads to better results in the classification phase in comparison with standard methods that do not use the recurrence during feature extraction.

  • [2] Maitra S., Chatterjee M., Sasidharan A., Sinha S., Mukhopadhyay K., (India) ,
    Working memory, impulsivity and emotional regulation correlates with frontal asymmetry of healthy young subjects during auditory session, pp. 365-378

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.024

    Abstract: Background: Specific frequency oscillations provide idea about functioning of underlying brain regions. Brain oscillations and event based assessment of cognitive functions like working memory (WM), impulsivity (Imp) and emotional regulation (ER) were reported to in uence each other in diflerent ethnic groups. But how these traits are regulated in healthy Indian adults was not explored widely. Aims: We analyzed link between scalp electrical activity and diflerent neuropsychological traits in higher education aspirants. Method: All the traits were self-assessed using standard questionnaires. QEEG was performed during an audio-sensory session. Tracings collected through BESS software were analyzed using SPSS. Results: Less impulsive individuals exhibited higher frontal theta and beta activity. Higher frontal theta activity was associated with higher ER, whereas higher theta and alpha activity showed association with WM deficit. Individuals with higher Imp and happiness exhibited higher frontal hemispheric asymmetry for theta and alpha, while those with lower asymmetry for alpha and beta activity showed higher ER. Beta asymmetry was positively related with happiness. Conclusions: We infer that variability in behaviour of healthy adults is in uenced by diflerential frontal brain impulses and could be considered for providing individualized assistance to emotionally vulnerable individuals.

  • [3] Brandejsky T. , (CZ)
    Versatile function in GPA, pp. 379-392

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.025

    Abstract: The paper, devoted to continuous versatile function application in the Genetic Programming Algorithm (GPA), begins with a discussion of similarities between GPA with versatile function and neural network. Then, the function set influence on GPA efficiency is discussed. In the next part, there is described a hybrid evolutionary algorithm that combines GPA for structure development and Evolutionary Strategy (ES) for parameters and constant optimization; which is herein much more significant than in the standard GPA. There is also discussed the setting of parameters of this hybrid algorithm and due to a diflerent function set. The original idea of a versatile function, which origins come from the area of fuzzy control systems, is formulated and explained. Four different implementations of this versatile function are discussed. On the base of experiments with the hybrid evolutionary algorithm providing symbolic regression of precomputed Lorenz attractor system data representing its dynamic behaviour; the comparison of three variants of versatile functions was formulated. The paper also presents ways how to set up hybrid evolutionary algorithm parameters like population sizes as well as limits of maximal population numbers for both algorithms: GPA for structural development and nested ES for parameters optimization. The versatile function concept is applicable but it requires the hybrid evolutionary algorithm use as it is explained in the paper.

  • [4] Contents volume 30 (2020), ... 393
  • [5] Authors index volume 30 (2020), ... 395

5/2020

  • [1] D. Húsek (CZ)
    Obituary – Professor Alexander A. Frolov passed away, pp. 281-282

      Full text    



  • [2] Koprivica M., (Serbia)
    Comparison of software packages for performing Bayesian inference, pp. 283-294

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.019

    Abstract: In this paper, we compare three state-of-the-art Python packages for Bayesian inference: JAGS , Stan and PyMC3. These packages are in focus because they are the most mature, and Python is among the most utilized programming languages for teaching mathematics and statistics in colleges . The experiment is based on real-world data collected for investigating the therapeutic touch nursing technique . It is analyzed through a hierarchical model with prior beta distribution and binomial likelihood function. The tools are compared by execution time and sample quality.

  • [3] Gulzat T., Lyazat N., Siladi V., Gulbakyt S., Maksatbek S., (Kazakhstan, SK) ,
    Research on predictive model based on classification with parameters of optimization, pp. 295-308

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.020

    Abstract: This paper effectively uses the data mining and optimization methods to investigate a classification based on decision trees algorithm, then optimizes by the method of grid search and cross-validation, which improves the prediction accuracy of the decision tree model for the PCs sales in practical application and solves insufficient training data, high computational cost, and low prediction accuracy. The main goal of the article is to predict PC sales using machine learning tools caused by various types of operating system factors in practical applications. This article proposes a combined predictive research model that fully reveals the benefits of optimization and neural networks, and also has a very accurate fit and forecasting accuracy. The proposed predictive model is implemented in the data science software platform RapidMiner. A decision tree model is executed, then the model's prediction capacity is evaluated and tested. Grid search optimizer is used to automatically build the final model using the best-optimized parameter for training the classifier. The paper combines grid the grid search and cross-validation to optimize the parameters of the decision tree to improve the classification prediction accuracy of the decision tree model. This article combines neural networks with optimization methods to establish a prediction model for laptop sales. This model gives full play to the advantages of optimization and neural networks and has very good fitting capabilities and prediction accuracy. Besides, the neural network for the prediction model has strong dynamic analysis capabilities. Once there are new observations, it can continue to be added to the modeling, which has high adaptability. The Neural Network algorithm has the highest accuracy of the predicted PC sales by evaluating the results of the five kinds of algorithms. The result for prediction accuracy shows the highest performance.

  • [4] Markošová M., Rudolf B., Náther P., Benušková L., (SK)
    Network models for changing degree distributions of functional brain networks, pp. 309-332

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.021

    Abstract: The purpose of this study was to investigate degree distributions of functional brain networks. Particular functional brain networks were constructed from the fMRI measurements of three groups of participants namely, young healthy participants, elderly healthy participants and elderly participants with Alzheimer disease. Functional brain networks were constructed for three different correlation thresholds of voxel activity correlated over time. We have noticed that the character of degree distribution changes when the value of correlation threshold decreases. In order to explain the degree distribution changes with the changes of value of correlation threshold, we created two different, yet related network models. The crucial factor both models contain is an increasing noise as the voxel activity correlation threshold is lowered, which in our models corresponds to an increase of the number of random correlations between the voxels { nodes of the functional network. The models account for how initially scale-free character of the degree distribution changes as the correlation threshold is lowered based on the processes of network growth and edge addition. The two models differ in the manner of preferential and random edge addition while the second model is a refinement of the first one. On average, the second model leads to a better quantitative match with the data. To our knowledge, such functional brain network models, which take into account the correlation threshold as an independent variable have not been introduced before.

  • [5] He. H., Yang X., Wu L., Wang G., (China)
    Iterated dilated convolutional neural networks for word segmentation, pp. 333-346

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.022

    Abstract: The latest development of neural word segmentation is governed by bi-directional Long Short-Term Memory Networks (Bi-LSTMs) that utilize Recurrent Neural Networks (RNNs) as standard sequence tagging models, resulting in expressive and accurate performance on large-scale dataset. However, RNNs are not adapted to fully exploit the parallelism capability of Graphics Processing Unit (GPU), limiting their computational efficiency in both learning and inferring phases. This paper proposes a novel approach adopting Iterated Dilated Convolutional Neural Networks (ID-CNNs) to supersede Bi-LSTMs for faster computation while retaining accuracy. Our implementation has achieved state-of-the-art result on SIGHAN Bakeoff 2005 datasets. Extensive experiments showed that our approach with ID-CNNs enables 3x training time speedups with no accuracy loss, achieving better accuracy compared to the prevailing Bi-LSTMs. Source code and corpora of this paper have been made publicly available on GitHub.


4/2020

  • [1] Qi X., Yuan Z., Han X., Liu S., (China)
    A discrete butterfly-inspired optimization algorithm for solving Permutation Flow-Shop scheduling Problems, pp. 211-230

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.015

    Abstract: Permutation Flow-Shop Scheduling Problem (PFSP) which exists in many manufacturing systems is a classic combinatorial optimization problem. Studies have shown that the PFSP including more than three machines belongs to the NP-hard problems and is difficult to solve. Based on a new bio-inspired algorithm - Artificial Buttery Optimization (ABO) algorithm, this paper presents a Discrete Artificial Buttery Optimization (DABO) algorithm to find the permutation that gives the smallest completion time or the smallest total ow time. The performance of the proposed algorithm is tested on well-known benchmark suites of Car, Reeves and Taillard. The experimental results show that the proposed algorithm is able to provide very promising and competitive results on most benchmark functions. The DABO algorithm is then employed for one production optimization problem.

  • [2] Zhou W., Liang Y.W., Ming Z., Dong H.B., (China)
    Earthquake prediction model based on danger theory in artificial immunity, pp. 231-247

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.016

    Abstract: Earthquake prediction is an extraordinarily stochastic process. Determining the occurrence time, location of epicenter and magnitude of a coming earthquake in the following month is an extremely difficult task. Nowadays, some geophysical, statistical and machine learning methods are adopted to predict earthquakes, however, for the insufficient medium-large seismic data, their results are not satisfactory. Due to there is no obvious empirical relationship between seismicity features, magnitude and location of a coming earthquake in a particular time window, an earthquake prediction approach based on danger theory is proposed in this paper. It extracts eight indicators calculated from earthquake data for recent years in Sichuan and surroundings by Gutenberg-Richter(GR) inverse power-law, and predicts quakes with magnitude lager than 4.5 during the following month by numerical differential based Dendritic Cell Algorithm (ndDCA). We compare this approach with six state-of-art earthquake prediction algorithms. Overall our algorithm yields the encouraging results in all the qualified parameters assessed, and it provides technical support for the application of earthquake prediction.

  • [3] Knobloch R., Mlýnek J., (CZ)
    Probabilistic analysis of the convergence of the differential evolution algorithm, pp. 249-263

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.017

    Abstract: Differential evolution algorithms represent an efficient framework to tackle complicated optimization problems with many variables and involved constraints. Nevertheless, the classic differential evolution algorithms in general do not ensure the convergence to the global minimum of the cost function. Therefore, the authors of the article designed a modification of these algorithms that guarantees the global convergence in the asymptotic and probabilistic sense. The modification consists in adding a certain ratio of random individuals to each generation formed by the algorithm. The random individuals limit the premature convergence to the local minimum and contribute to more thorough exploration of the search space. This article concentrates specifically on the role of random individuals in the identification of the global minimum of the cost function. Besides, the paper also contains some useful estimates of the probability of finding the global minimum of the corresponding cost function.

  • [4] Li W., Li B., Guo H.L., Fang Y.X., Qiao F.J., Zhou S.W., (China)
    The ECG signal classification based on ensemble learning of PSO-ELM algorithm, pp. 265-279

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.018

    Abstract: ECG anomaly detection plays an important role in clinical medicine. So far, a number of ECG recognition technologies have emerged in this field, but most often suffer from slow training and instability. Considering that the Extreme Learning Machine (ELM) and Particle Swarm Optimization (PSO) algorithm have the advantages of fast learning speed and strong generalization ability, this paper integrates multiple independent PSO-ELM model and proposes a novel ensemble learning framework termed as E-PSO-ELM to realize ECG signals recognition. More specifically, the individual PSO-ELM adopts the input weight and hidden layer deviation of ELM as the particles in the PSO algorithm, and takes the root mean square error of ELM training sample as the adaptive value of the particles, so as to enhance the stability of the network and realize high ECG recognition rate. The simulation results on MIT-BIH Arrhythmia Database show that E-PSO-ELM has a high classification accuracy rate of 98.23 %. In addition, compared with other algorithms, the stability of E-PSO-ELM is more prominent, which can reduce the probability of operating errors. Therefore, E-PSO-ELM has a high practical value.


3/2020

  • [1] Gao W., Fang Y., Zhang F., Yang Z., (China)
    Representation learning of knowledge graphs using convolutional neural networks, pp. 145-160

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.011

    Abstract: Knowledge graphs have been playing an important role in many Artificial Intelligence (AI) applications such as entity linking, question answering and so forth. However, most of previous studies focused on the symbolic representation of knowledge graphs with structural information, which cannot deal well with new entities or rare entities with little relevant knowledge. In this paper, we propose a new deep knowledge representation architecture that jointly encodes both structure and textual information. We first propose a novel neural model to encode the text descriptions of entities based on Convolutional Neural Networks (CNN). Secondly, an attention mechanism is applied to capture the valuable information from these descriptions. Then we introduce position vectors as supplementary information. Finally, a gate mechanism is designed to integrate representations of structure and text into the joint representation. Experimental results on two datasets show that our models obtain state-of-the-art results on link prediction and triplet classification tasks, and achieve the best performance on the relation classification task.

  • [2] Zhou Z., Fu Y., Zhao J. (China)
    An efficient method for surface reconstruction based on local coordinate system transform and partition of unity , pp. 161-176

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.012

    Abstract: Radial basis function (RBF) has been extensively applied for surface reconstruction from scattered 3D point data due to its strong ability of approx- imation. However, additional information, such as off-surface points, are usually required to be appended into constraints for determining the parameters, which apparently increases the computation cost and data unreliability. To avoid adding additional off surface point constraints, a novel surface reconstruction approach based on local coordinate system transform and partition of unity is proposed in this paper. Firstly, the explicit RBF functions are constructed to approximate the local surface patches, and then it is transformed into an equivalent implicit surface reconstruction form by local system coordinate transformation. Compared with the local implicit surface approximation, the proposed local explicit surface approxima- tion method is capable of avoiding trivial solution occurred in RBF approximating, and does not increase the scale of data solution. A number of comparison exper- iments of the proposed method with the traditional RBF-based method and the multi-level partition of unity (MPU) method are carried out on some kinds of large dataset, non-uniformity dataset, noisy dataset. The experimental results illustrate that the proposed method is robust and effective in dealing with large-scale point clouds surface reconstruction.

  • [3] Naseem M.T., Qureshi I.M., Atta-ur-Rahman, Muzaffar M.Z., (Pakistan, Saudi Arabia) ,
    Robust and fragile watermarking for medical images using redundant residue number system and chaos, pp. 177-192

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.013

    Abstract: This research discusses a novel watermarking scheme using redundant residue number system and chaos. The salient feature of said research is that image remains fragile while the watermark information is made robust. Image pixels are converted into residues so that the unaided eye could not see the image contents. To make the image invisible to the unaided eye, only the ROI part of image is passed through the Residue Number System thus, to enhance the secrecy of the image. While converting the ROI part of image into residues, there are some residues which exceed eight bits so, these residues are converted to exact eight bits by pertaining some intelligent mechanism. To achieve the robustness of watermark, firstly redundant residues of watermark are made and then the resultant watermark is encoded through error correcting codes. To achieve the fragility of image, hashing technique is utilized. Hash of the entire image but with the residued ROI is combined with the encoded and redundant residued watermark and then resultant watermark is embedded in the Region of non-interest (RONI) zone of native image rooted on the chaotic key in order to enhance the security of the watermark. In case of no tampering, fragile watermark can be successfully recovered as well as exact recovery of the original image but if the image is attacked, the fragile watermark is destroyed while the robust watermark is extracted with better readability.

  • [4] Provinský P., (CZ)
    Floppy logic as a generalization of standard Boolean logic, pp. 193-209

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.014

    Abstract: The topic of this article is a floppy logic, a new multi-valued logic. Floppy logic is related to fuzzy logic and the theory of probability, but it also has interesting links to probability logic and standard Boolean logic. It provides a consistent and simple theory that is easy to apply in practice. This article examines the isomorphism theorem, which plays an important role in floppy logic. The theorem is described and proved. The most important consequences of the isomorphism theorem are: 1) All statements which are equivalent in standard Boolean logic are also equivalent in floppy logic. 2) Floppy logic has all the properties of standard Boolean logic which can be formulated as an equivalence. These include, for example, distributivity, the contradiction law, the law of excluded middle, and others. The article mainly examines floppy implication. We show that floppy implication does not satisfy Adam's Thesis and that floppy logic is not limited by Lewis' triviality result. We also present a range of inference rules which are generalizations of modus ponens and modus tollens. These rules hold in floppy logic, and of course, also apply to standard Boolean logic. All these results lead us to the notion that floppy logic is a many-valued generalization of standard Boolean logic.


2/2020

  • [1] Bouchner P. (CZ)
    EDITORIAL – Prof. Ing. Mirko Novák, DrSc. passed away , pp. 77-84

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.006



  • [2] Wu H., Song Q., Jin G. (China)
    Underwater acoustic signal analysis: preprocessing and classification by deep learning , pp. 85-96

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.007

    Abstract: The identification and classification is important parts of the research in the field like underwater acoustic signal processing. Recently, deep learning technology has been utilized to achieve good performance in the underwater acoustic signal case. On the other side, there are still some problems should be solved. The first one is that it cannot achieve high accuracy by the dataset that is transformed into audio spectrum. The second one is that the accuracy of classification on the dataset is still low, so that, it cannot satisfy the real demand. To solve those problems, we firstly evaluated four popular spectrums (Audio Spectrum, Image Histogram, Demon and LOFAR) for data preprocessing and selected the best one that is suitable for the neural networks (LeNet, ALEXNET, VGG16). Then, among these methods, we modified a neural network(LeNet) to fit the dataset that is transformed by the spectrum to improve the classification accuracy. The experimental result shows that the accuracy of our method can achieve 97.22 %, which is higher than existing methods and it met the expected target of practical application.

  • [3] Sekeroglu B., Dimililer K., (Turkey)
    Review and analysis of hidden neuron number effect of shallow backpropagation neural networks, pp. 97-112

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.008

    Abstract: Shallow neural network implementations are still popular for real-life classification problems that require rapid achievements with limited data. Parameters selection such as hidden neuron number, learning rate and momentum factor of neural networks are the main challenges that causes time loss during these implementations. In these parameters, the determination of hidden neuron numbers is the main drawback that affects both training and generalization phases of any neural system for learning efficiency and system accuracy. In this study, several experiments are performed in order to observe the effect of hidden neuron number of 3-layered backpropagation neural network on the generalization rate of classification problems using both numerical datasets and image databases. Experiments are performed by considering the increasing number of total processing elements, and various numbers of hidden neurons are used during the training. The results of each hidden neuron number are analyzed according to the accuracy rates and iteration numbers during the convergence. Results show that the effect of the hidden neuron numbers mainly depends on the number of training patterns. Also obtained results suggest intervals of hidden neuron numbers for different number of total processing elements and training patterns.

  • [4] Zhang Y., Jin S., Wu Y., Zhao T., Yan Y., Li Z., Li Y., (China)
    A new intelligent supermarket security system, pp. 113-131

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.009

    Abstract: With the rapid development of artificial intelligence in recent years, the application of intelligent security has become increasingly widespread. This paper presents a new intelligent system that uses Convolutional Neural Network (CNN) combined with a high-resolution camera to identify the theft behavior of customers. The CNN extracts relevant information from the theft and non-theft behavior of customers in supermarkets to establish a recognition model. Our results show that, by updating the data sets, the recognition model can be continuously optimized, and the average recognition accuracy finally reaches 83 %. The proposed system can independently identify the theft and non-theft behavior in video surveillance and sound alarm on the theft behavior in time. The advantages of the system are its low cost and high precision, which show excellent commercial value and application prospects.

  • [5] Svítek M. (CZ)
    Quantum informatics and soft systems modeling, pp. 133-144

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.010

    Abstract: This paper elaborates on the area of physical information analogies and introduces new features such as the distance between wave probabilistic functions or sets of new information quantities such as strength, strength moment, strength potential energy and generalized charge. New parameters are used to define the rules for a quantum node. The knowledge cycle, which is equivalent to the Otto thermodynamic cycle, is adopted for modeling of the soft systems together with its static and dynamic information stability. Looking at the closed knowledge cycle, the evolutionary field equivalent to a magnetic field is therefore determined.


1/2020

  • [1] Gitto L., Massini G., Mennini F.S., Mento C., Buscema P.M. (Italy)
    Affective symptoms and postural abnormalities as predictors of headache: an application of artificial neural networks , pp. 1-26

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.001

    Abstract: Chronic headache is a major liability in the individuals' quality of life. Identifying in advance the main features common to patients with headache may allow planning a preventive strategy of intervention. An artificial neural network model (Auto Contractive Maps { AutoCM), aimed at analyzing the correlations between patients' characteristics, affective symptoms and posture indicators has been developed in this paper. Patients suffering from chronic headache were observed at a neurological centre in Sicily (Italy). Headache and affective states were measured using the Profile of Mood States (POMS), the Beck Depression Inventory (BDI), the Toronto Alexithymia Scale (TAS-20) and the Repression Scale. Postural evaluations were carried through a stabilometric platform. The method of analysis selected allowed to reconstruct some records that were missing, through a Recirculation AutoAssociative Neural Network, and to obtain sound results. The results showed how some items from TAS-20, Repression and POMS were closely linked. The postural abnormalities were correlated primarily with repression features. The highest scores of the POMS were correlated with the items of the BDI. The results obtained lead to interesting remarks about the common traits to patients with headache. The main conclusion lies in the potentialities offered by the new methodology applied, that may contribute, overall, to a better understanding of the complexity of chronic diseases, where many factors concur to define patients' health conditions.

  • [2] Hu J., Cao S., Xu C., Yao J., Xie Z. (China)
    High-accuracy motion control of a motor servo system with dead-zone based on a single hidden layer neural network , pp. 27-44

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.002

    Abstract: There always exist parametric uncertainties, bounded disturbances and some other unknown nonlinearities such as the input dead-zone in physical motor servo systems, which can degrade the system's control performance. In this paper, a composite control strategy is proposed for high-accuracy motion control of a torque-controlled motor servo system with dead-zone. A smooth and continuous mathematical model is used to provide an approximate inverse transformation of the input-output dead-zone needed for feedback linearization. A single-layer neural network capable of on-line learning is designed to compensate for the inversion error, which comes from the approximate inversion. A stable weights adaption law for the on-line neural network is derived. In addition, a parameter adaptation law is also derived for handling the parametric uncertainty, and a nonlinear robust feedback term is designed to inhibit the in uence of the imperfect modeling, compensation error or other disturbances. Lyapunov theorem is used to prove the stability of the proposed control algorithm with the weights and parameters adaptation law. Extensive comparative simulation results are used to illustrate the effectiveness and advancement of the proposed controller compared with several other main-stream controllers.

  • [3] Moos P., Novák M., Votruba Z. (CZ)
    Parametric sensitivity in decision making process, pp. 45-53

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.003

    Abstract: This paper introduces a possibility of application of parametric sensitivity appearing in processes of decision making in systems represented by the general production functions of the Tinbergen type, depending on information content, information flow and qualification of human resources. The so-called parametric sensitivity considering the information content I as an ordering parameter, dependent on the information flow , applied on production function. The theory of production function describes the relation between physical outputs of a production process and physical inputs, i.e. factors of production. Finally, the influence of knowledge in information content I, leading to correct decision, is demonstrated through the parametric sensitivity concept. For this invention, J. Tinbergen and R. Frisch achieved in 1969 the "Nobel price of the Swedish National Bank". Besides, the production functions theory surprisingly represents also a tool for finding the reasons of living bodies behavior.

  • [4] Svítek M. (CZ)
    Wave composition rules in quantum system theory, pp. 55-64

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.004

    Abstract: The paper presents the new approach to wave composition rules for advanced modeling of soft systems in quantum system theory. Firstly, the interpretation of phase parameters is given. The phase parameters are essential to specify the mathematical operations assigned to different relations among subsystems, e.g. co-operation, connection, co-existence, competition. Using wave composition rules, we are able to create more complex and sophisticated quantum circuits. We present the application of methodology on three illustrative examples.

  • [5] Dembani R., Zheng W., Sun M., Nooruddin (China)
    Unsupervised facial expression detection using genetic algorithm, pp. 65-75

      Full text     DOI: http://dx.doi.org/10.14311/NNW.2020.30.005

    Abstract: Interpersonal communication can be done by understanding the clues of facial expressions. As its importance increase in behavior and clinical studies, so automatic detection of facial expressions is an open research area for the last few decades. Efforts of expression detection by a human being are easy and effective but the machine needs some more understanding. This paper proposes a face expression clustering using a genetic algorithm. Image get convert into binary format for finding the related cluster selection in different phases of genetic algorithm. Proposed work has utilized a modified teacher learning-based optimization algorithm where the population gets updated in each phase to get the best representative features. A real dataset of facial expression was used in this work. A comparison of the proposed model was done with existing models on different evaluation parameters. It was obtained that the proposed work has improved precision, recall, the accuracy of facial expression identification without any training.