Neural Networks and Neuro-Fuzzy Syetems

Industrial control systems are nowadays exposed in environments with rapid and unstable parameter changes and uses measuring equipments with critical output sensitivity. In the case of thermal gas analyzer, measurement errors are contributed by temperature, gas flow, and pressure. Error compensation is a key problem for these control systems. In recent years, it has been proven in the literature that artificial neural network (ANN) is a reliable and low cost solution to manage errors. Among all the algorithms of ANN, the back propagation is commonly used because of its simplicity and learning methodology is easy to realize. However, it has two notable drawbacks: (a) it is likely to run into local minimum, and (b) convergence is slow. Thermal conductivity gas analyzer often works in adverse surroundings, which requires fast and accurate measurements. Therefore, a strong learning network is needed. This paper proposes a novel thermal gas analyzer using adaptive neuro-fuzzy inference system. The effectiveness and validity of the proposed method is verified by simulation studies using MATLAB. Fuzzy membership rules are created to allow regulation of learning parameters. Further, the fuzzy adaptive network model is constructed to train large data samples while the high precision compensation of sensor error is realized by the improved flow. Simulation results reveal that the convergence speed and output accuracy is improved and the learning parameters in thermal gas analyzer are automatically corrected by the proposed method in comparison with the back propagation algorithm of artificial neural network.

Intrusion detection has attracted a considerable interest from researchers and industries. The community, after many years of research, still faces the problem of building reliable and efficient IDS that are capable of handling large quantities of data, with changing patterns in real time situations. The work presented in this manuscript classifies intrusion detection systems (IDS). Moreover, a taxonomy and survey of shallow and deep networks intrusion detection systems is presented based on previous and current works. This taxonomy and survey reviews machine learning techniques and their performance in detecting anomalies. Feature selection which influences the effectiveness of machine learning (ML) IDS is discussed to explain the role of feature selection in the classification and training phase of ML IDS. Finally, a discussion of the false and true positive alarm rates is presented to help researchers model reliable and efficient machine learning based intrusion detection systems.

Adaptive neuro-fuzzy inference system (ANFIS) is efficient estimation model not only among neuro-fuzzy systems but also various other machine learning techniques. Despite acceptance among researchers, ANFIS suffers from limitations that halt applications in problems with large inputs; such as, curse of dimensionality and computational expense. Various approaches have been proposed in literature to overcome such shortcomings, however, there exists a considerable room of improvement. This paper reports approaches from literature that reduce computational complexity by architectural modifications as well as efficient training procedures. Moreover, as potential future directions, this paper also proposes conceptual solutions to the limitations highlighted.

Scope of the book:
This book focusses on the technical concepts of deep learning and its associated branch Neural Networks for the various dimensions of image processing applications. The proposed volume intends to bring together researchers to report the latest results or progress in the development of the above-mentioned areas. Since there is a deficit of books on this specific subject matter, the editors aim to provide a common platform for researchers working in this area to exhibit their novel findings.
Topics of Interest:
This book solicits contributions, which include the fundamentals in the field of Deep Artificial Neural Networks and Image Processing supported by case studies and practical examples. Each chapter is expected to be self-contained and to cover an in-depth analysis of real life applications of neural networks to image analysis.

The diagnosis of blood related diseases involves the identification and characterization of a patient's blood sample. As such, automated methods for detecting and classifying the types of blood cells have important medical applications in this field. Although deep convolutional neural network (CNN) and the traditional machine learning methods have shown good results in the classification of blood cell images. Blood cells mainly include red blood cells, white blood cells and platelets. In blood, leucocyte plays an important role in the human immune function, so it is also called the immune cell. Usually, hematologists use granulated information and shape information in leukocytes to divide white blood cells into granular cells: neutrophil, eosinophil, basophil and non-granular cells: monocyte and lymphocyte. We apply the transfer learning method to transfer the weight parameters that were pre-trained on the Image Net dataset to the CNN section, and adopted a custom loss function to allow our network to train and converge faster and with more accurate weight parameters. Hence this process is applied to detect the abnormality of the cell by means of segmentation and classification. The classification results have shown an accuracy of 88.04% for KNN and 54% for ANN. The proposed system work can be further enhanced by taking other classifiers.

This paper considers the problem of oscillations in a synchronous generator connected to infinite bus through transmission lines. Two on-line control techniques, namely, artificial neural networks (ANN) and simulated annealing (SA) are utilized to cancel the oscillations in synchronous generators (SG). Simulation results of applying external disturbances to the synchronous generator controlled by the proposed simulated annealing controllers are compared to results obtained by using neural network controllers. These control schemes contribute to preventing system instability by suppressing the low-frequency oscillations arising from power grid fault disturbances. The proposed on-line SA and NN integrate a voltage regulator and a power system stabilizer to obtain near-optimal solutions of the problem through utilizing functions evaluation. They can be adopted to replace the conventional automatic voltage regulator (AVR) with power system stabilizer (PSS) of the generator. Simulation results show that algorithms can efficiently and effectively solve such optimization problems within short time. In addition, they are presented to demonstrate the effectiveness and advantage of the control system of synchronous generator (SG) in comparison with the conventional control scheme so as to allow the generator to operate closely to its steady state stability limits.

Multilabel Image Tagging is one of the most important challenges in computer vision with many real world applications and thus we have used Deep Neural Networks for Image Annotation to boost performance. This experiment is performed on NUS-WIDE Dataset with 1K Tags. I. INTRODUCTION Multilabel image annotation is an important and challenging problem in computer vision. Most existing work focus on single-label classification problems, where each image is assumed to have only one class label. However, this is not necessarily true for real world applications, as an image may be associated with multiple tags. As a practical example, images from Flickr are used as an example which has multiple tags, such as objects, activities, and scene descriptions. Images on the Internet, in general, are usually associated with sentences or descriptions, instead of a single class label, which is a type of multitagging. Therefore, it is a practical and important problem to accurately assign multiple labels to one image. Single-label image classification has been extensively studied in the vision community, the most recent advances reported on the large-scale ImageNet Architecture. Most existing work focus on designing visual features for improving recognition accuracy. For example, sparse coding, Fisher vectors, and VLAD have been proposed to reduce the quantization error of " bag of words "-type features. Very recently, deep convolutional neural networks (CNN) have demonstrated promising results for single-label image classification. Such algorithms have focused on one vs all, but one have worked on the multilabel image annotation problem. In this work, we used the highly expressive convolutional network for the problem of multilabel image annotation. We employed a similar network structure to as used in Image Net, which contains several convolutional and dense connected layers as the basic architecture. We studied and compared several other popular multilabel losses, such as the ranking loss that optimizes the area under ROC curve (AUC), and the cross-entropy loss used in Tagprop. Specifically, we propose to use the top-k ranking loss, inspired by, for embedding to train the network. Using the largest publicly available multilabel dataset NUS-WIDE, we observe a significant performance boost over conventional features, reporting the best retrieval performance. We even performed test on Triplet Loss giving much better results. II. MOTIVATION Multilabel image annotation is an important and challenging problem in computer vision. Most existing work focus on single-label classification problems, where each image is assumed to have only one class label. However, this is not necessarily true for real world applications, as an image may be associated with multiple tags. As a practical example, images from Flickr are used as an example which has multiple tags, such as objects, activities, and scene descriptions. Images on the Internet, in general, are usually associated with sentences or descriptions, instead of a single class label, which is a type of multitagging. Therefore, it is a practical and important problem to accurately assign multiple labels to one image.

— Modeling time series is often associated with the process forecasts certain characteristics in the next period. One of the methods forecasts that developed nowadays is using artificial neural network or more popularly known as aneural network. Use neural network in forecasts time series can be agood solution, but the problem is network architecture and the training method in the right direction. General Regression Neural Network (GRNN) is one of the network model radial basis that used to approach a function. GRNN including model neural network model with a solution that quickly, because it is not needed each iteration in the estimation weight. This model has a network architecture that wasa number of units in pattern layer in accordance with the number of input data. One of the application GRNN is to predict the crude oil by using a model GRNN.From the training and testing on the data obtained by the RMSE testing 1.9355 and RMSE training 1.1048.Model is good to be used to give aprediction that is quite accurate information that is shown by the close target with the output

In this paper, a Fuzzy Neural Petri Net (FNPN) controller has been designed established on Particle Swarm Optimization (PSO) for controlling the path tracking of Wheeled Mobile Robot (WMR). The path planning controller problem has been solved using two FNPN controllers to get the desired velocity and azimuth. The PSO method has used to detection the optimal values parameters of FNPN controllers. The overall models of wheeled mobile robot for path tracking control created on PSO algorithm are implemented in Simulink-Matlab. Simulation outcomes demonstrate the suggested FNPN controllers is more effectiveness and has good dynamic performance than the conventional methods.

Automatic recognition of cardiac arrhythmias is important for diagnosis of cardiac abnormalies. Several algorithms have been proposed to classify ECG arrhythmias; however, they cannot perform very well. Therefore, in this paper, an expert system for ElectroCardioGram (ECG) arrhythmia classification is proposed. Discrete wavelet transform is used for processing ECG recordings, and extracting some features, and the Multi-Layer Perceptron (MLP) neural network performs the classification task. Two types of arrhythmias can be detected by the proposed system. Some recordings of the MIT-BIH arrhythmias database have been used for training and testing our neural network based classifier. The simulation results show that the classification accuracy of our algorithm is 96.5% using 10 files including normal and two arrhythmias.

This paper focuses on the study of short term load forecasting (STELF) using interval Type-2 Fuzzy Logic (IT2FL) and feed-forward Neural Network with back-propagation (NN-BP) tuning algorithm to improve their approximation capability, flexibility and adaptiveness. IT2FL for STELF is presented which provides additional degrees of freedom for handling more uncertainties for improving prediction accuracy and reducing cost. The IT2FL comprises five components which include; the fuzzification unit, the knowledge base, the inference engine, the type reducer and the defuzzification unit. Gaussian membership function is used to show the degree of membership of the input variables. The lower and upper membership functions (fired rules) as well as their consequent coefficients of IT2FL are fed into a (NN) which produces a crisp value coresponding to the optimal defuzzified output of IT2FLSs. The NN type reducer is trained to optimize parameters of membership function (MF) so as to produce an output with minimum error function with the purpose of improving forecasting performance of IT2FLS models. The IT2FNN system has the ability to overcome the limitations of individual technique and enhances their strengths to handle electric load forecasting. The IT2FNN is applied for STELF in Akwa Ibom State-Nigeria. The result of performance of IT2FNN is compared with IT2FLS and T1FLS methods for short term load forecasting with MSE of 0.00123, 0.00185 and 0.00247 respectively. Also, the results of forecasting are compared using RMSE of 0.035, 0.043 and 0.035 respectively, indicating a best accurate forecasting with IT2FNN. In addition, the result of performance of IT2FNN is compared with IT2FLS and T1FLS methods for short term load forecasting with MAPE of 1.5%, 3% and 4.5% respectively. Simulation results show that the IT2FNN approach takes advantages of accuracy and efficiency and performs better in prediction than IT2FL and T1FL methods in power load forecasting task.. KEYWORDS Interval type-2 fuzzy logic; feed-forward neural network; back propagation; electric load; Interval type-2 fuzzy neural networks; type-1 fuzzy logic.

In this paper, we present a soft computing modelof decision support systems for diagnosing diseases andprescribing herbal prescriptions by oriental medicine.Inputs to the model are severities of observed symptoms onpatients and outputs from the model are a diagnosis of disease states and corresponding herbal prescriptions.First, having used fuzzy inferences, the most seriousdisease state in which the patient appears to be infected isdetermined. Next, an herbal prescription written insuitable herbs with reasonable amounts for treating thatdisease state is given by neural networks. Finally, wedescribe an application of this model in rheumatismdiagnoses and show its evaluations

The growth of Internet and other web technologies requires the development of new algorithms and architectures for parallel and distributed computing. International journal of Distributed and parallel systems is a bi monthly open access peer-reviewed journal aims to publish high quality scientific papers arising from original research and development from the international community in the areas of parallel and distributed systems. IJDPS serves as a platform for engineers and researchers to present new ideas and system technology, with an interactive and friendly, but strongly professional atmosphere.

In this paper a new interval Type-2 fuzzy neural network will be presented for function approximation. The proposed neural network is based on Locally Linear Model Tree (LOLIMOT) which is a fast learning algorithm for Locally Linear Neuro-Fuzzy Models (LLNFM). In this research, main measures are to be robust in presence of outlier data and be fast in refining steps. The proposed combination between LOLIMOT learning algorithm and interval type 2 fuzzy logic systems presents a good performance both in robustness and speed measures. The results show that the proposed method has good robustness in presence of noise as we can see in experiments conducted using corrupted data. Also this method has eligible speed as it can be seen in the results.

This paper focuses on the study of short term load forecasting (STELF) using interval Type-2 Fuzzy Logic (IT2FL) and feed-forward Neural Network with back-propagation (NN-BP) tuning algorithm to improve their approximation capability, flexibility and adaptiveness. IT2FL for STELF is presented which provides additional degrees of freedom for handling more uncertainties for improving prediction accuracy and reducing cost. The IT2FL comprises five components which include; the fuzzification unit, the knowledge base, the inference engine, the type reducer and the defuzzification unit. Gaussian membership function is used to show the degree of membership of the input variables. The lower and upper membership functions (fired rules) as well as their consequent coefficients of IT2FL are fed into a (NN) which produces a crisp value coresponding to the optimal defuzzified output of IT2FLSs. The NN type reducer is trained to optimize parameters of membership function (MF) so as to produce an output with minimum error function with the purpose of improving forecasting performance of IT2FLS models. The IT2FNN system has the ability to overcome the limitations of individual technique and enhances their strengths to handle electric load forecasting. The IT2FNN is applied for STELF in Akwa Ibom State-Nigeria. The result of performance of IT2FNN is compared with IT2FLS and T1FLS methods for short term load forecasting with MSE of 0.00123, 0.00185 and 0.00247 respectively. Also, the results of forecasting are compared using RMSE of 0.035, 0.043 and 0.035 respectively, indicating a best accurate forecasting with IT2FNN. In addition, the result of performance of IT2FNN is compared with IT2FLS and T1FLS methods for short term load forecasting with MAPE of 1.5%, 3% and 4.5% respectively. Simulation results show that the IT2FNN approach takes advantages of accuracy and efficiency and performs better in prediction than IT2FL and T1FL methods in power load forecasting task.. KEYWORDS Interval type-2 fuzzy logic; feed-forward neural network; back propagation; electric load; Interval type-2 fuzzy neural networks; type-1 fuzzy logic.

Deep neural network as a part of deep learning algorithm is a state-of-the-art approach to find higher level representations of input data which has been introduced to many practical and challenging learning problems successfully. The primary goal of deep learning is to use large data to help solving a given task on machine learning. We propose an methodology for image de-noising project defined by this model and conduct training a large image database to get the experimental output. The result shows the robustness and efficient our our algorithm.

Modeling time series is often associated with the process forecasts certain characteristics in the next period. One of the methods forecasts that developed nowadays is using artificial neural network or more popularly known as a neural network. Use neural network in forecasts time series can be a good solution, but the problem is network architecture and the training method in the right direction. One of the choices that might be using a genetic algorithm. A genetic algorithm is a search algorithm stochastic resonance based on how it works by the mechanisms of natural selection and genetic variation that aims to find a solution to a problem. This algorithm can be used as teaching methods in train models are sent back propagation neural network. The application genetic algorithm and neural network for divination time series aim to get the weight optimum. From the training and testing on the data index share price euro 50 obtained by the RMSE testing 27.8744 and 39.2852 RMSE training. The weight or parameters that produced by has reached an optimum level in second-generation 1000 with the best fitness and the average 0.027771 the fitness of 0.0027847.Model is good to be used to give a prediction that is quite accurate information that is shown by the close target with the output.

Every day, the estimated volume of data which is generated per day is 2.6 quintillion bytes. From the last two years, there is a lot of data generation and execution is taking rise due to feasible technologies and devices. To make the information accessible with ease, we need to classify the information data and predict an accurate or at least an approximate expected result which is forwarded to the end user client. To achieve the said process, the information technology industries are more concerned with machine learning and edge computing. Machine learning is a integral subset of artificial intelligence .In machine learning, the foremost step towards achieving the above task is to observe the data which is produced in large amount, later classify the data to make the system learn (train) from the old data (experience) that is stored at the server level and finally predict an estimation as a result. The obtained result is been transformed onto the devices which have made a request for a particular data. These devices are remotely located at the corner of the central data center. The process in which the execution of the information data is done at the corner of the data center is called as edge computing. In today's world of high computation, these two technologies i.e machine learning and edge computing are creating an overwhelming significance for its usage in the business market and end user clients. Here, we try to explain few possibilities of integrating the two technologies.

Scope International Journal of Artificial Intelligence and Soft Computing (IJAISC) is a open access journal that publishes articles which contribute new results in all areas of Computer Networks & Communications. The journal focuses on all technical and practical aspects of Computer Networks & data Communications. Authors are solicited to contribute to this journal by submitting articles that illustrate research results, projects, surveying works and industrial experiences that describe significant advances in the Computer Networks & Communications. Topics of interest include, but are not limited to, the following

The suggested method helps predicting vehicles movement in order to give the driver more time to react and avoid collisions on roads. The algorithm is dynamically modelling the road scene around the vehicle based on the data from the on-board camera. All moving objects are monitored and represented by the dynamic model on a 2D map. After analyzing every object's movement, the algorithm predicts its possible behavior.

Wireless sensor networks (WSNs) are a community of large-scale, low-power, low-cost wireless sensor nodes. This paper proposes a new fuzzy-neural based routing protocol, called Routing Protocol using Fuzzy system and Neural node, RPFN. Data packets are transferred from sensor nodes to a desired base station by hop-to-hop delivery. When a sensor node has a new sensed data or a data packet has been received from its neighbors, it selects an appropriate neighbor called candidate node by a fuzzy inference system and a neural node. The proposed Perceptron-based neural node uses four essential parameters including remaining energy, distance to the base station, available buffer, and link quality to choose the best candidate node according to local information. Moreover, parameter “link quality” is determined by the proposed fuzzy system based on distance to neighbor node and response rate. Simulation results demonstrate that RPFN surpasses some existing routing protocols in terms of packet delivery ratio and network lifetime.

In this paper we have used 3D motion capture data with the aim of detecting and classifying specific human actions. In addition to recognition of basic action classes, actor styles and characteristics such as gender, age, and energy level have also been subject to classification. We have applied and compared three main methods: nearest neighbour search, hidden Markov models, and artificial neural networks. Using
these techniques, we have proposed exhaustive algorithms for detection of actions in a motion piece and subsequently classifying the segmented actions and respective characteristics of the actors. We have tested the methods for various sequences and compared the results for a comprehensive evaluation of each of the proposed techniques. Our findings can be largely used for general classification of human motion data for multimedia applications as well as sorting and classifying data sets of human motion data especially those acquired using visual marker-based motion capture systems such as the one employed in this research.

New research indicates that the “living force” can be equated to the overall matter/energy pattern of an organism, while mind and consciousness are the electric and magnetic counterparts of the matter/energy pattern. In fact, all living organisms are complex material/energy patterns (comprising the structural and biochemical body) supplemented by extremely complex electric (electromagnetic scalar potential) field patterns and magnetic (electromagnetic vector potential) field patterns, constituting mind and consciousness. The brain alone should no longer be considered the home of mind and consciousness in the body. It only seems so because of the dense concentration of neurons in the brain. A strict study of the magnetic vector potential field patterns associated with neurons and neural nets demonstrates how thoughts and streams of thought originate in the brain and are stored magnetically at and within the points in space-time occupied by the brain and body. Microtubules, which act as bio-magnetic induction coils, form electromagnetic transceivers in conjunction with axon wall capacitors within neurons. They are the primary structural bio-unit used for building, storing and retrieving memories in the mind. Mind and consciousness could thus be considered a four-dimensional electric-magnetic potential field holographic body that maintains a semi-independent existence and survives death.

Due to the increasing deployment of vehicles in human societies and the necessity for smart traffic control, anomaly detection is among the various tasks widely employed in traffic monitoring. As the issue of urban traffic and their relative smart monitoring systems have gained popularity among researchers in recent years, there exist several studies in this regard. In most of these studies, classification is performed based on the behavior of drivers, where a set of default trajectories are used in order to learn the system and classify the related data. However, two under-studied challenges are the lack of access to sufficient data to provide an efficient model, along with the lack of access to anomaly data that covers all possible abnormal trajectories. While the former challenge can be tackled through long-term data recording, the latter requires appropriate considerations. To this aim, we have utilized a combination of optimized convolutional neural network and fuzzy neural network classifiers, along with autoencoding neural networks. The final combination occurs at the decision level. First, the CNN-ANFIS classifier assigns the input trajectory to one of the predefined categories. Then, the trained autoencoder networks examine the result in order to find whether the trajectory is normal or abnormal. Obtaining 87.5% accuracy on QMUL and 99.5% on the T15 datasets confirms the superior performance of the proposed method.

We have used neural network water level trend prediction (NNWLTP) in support of a water level sensing project. The NNWLTP approach allows dynamic change in water level sampling frequency, which will reduce power consumption and extend battery life in energy constrained devices. This paper deals primarily with the NNWLTP, which would allow sampling frequency change commands to be transmitted to

Artificial Neural Networks (ANNs) have been successfully used in Pattern Recognition tasks. Evolutionary Spiking Neural Networks (ESNNs) constitute an approach to design third-generation ANNs (also known as Spiking Neural Networks, SNNs) involving Evolutionary Algorithms (EAs) to govern some intrinsic aspects of the networks, such as topology, connections and/or parameters. Concerning the practicality of the networks, a rather simple standard is commonly used; restricted feed-forward fully-connected network topologies deprived from more complex connections are usually considered. Notwithstanding, a wider prospect of configurations in contrast to standard network topologies is available for research. In this paper, ESNNs are evolved to solve pattern classification tasks, using an EA-based algorithm known as Grammatical Evolution (GE). Experiments demonstrate competitive results and a distinctive variety of network designs when compared to a more traditional approach to design ESNNs.

Neural networks and genetic algorithms have been in the past successfully applied, separately, to controller tuning problems. In this paper we purpose to combine its joint use, by exploiting the nonlinear mapping capabilities of neural networks to model objective functions, and use them to supply their values to a genetic algorithm which performs on-line minimization. Simulation results show that this is a valid approach, offering desired properties for on-line use such as a dramatic reduction in computation time and avoiding the need of perturbing the closed-loop operation.

Economic indicators such as Consumer Price Index (CPI) have frequently used in predicting future economic wealth for financial policy makers of respective country. Most central banks, on guidelines of research studies, have recently adopted an inflation targeting monetary policy regime, which accounts for high requirement for effective prediction model of consumer price index. However, prediction accuracy by numerous studies is still low, which raises a need for improvement. This manuscript presents findings of study that use neuro fuzzy technique to design a machine-learning model that train and test data to predict a univariate time series CPI. The study establishes a matrix of monthly CPI data from secondary data source of Tanzania National Bureau of Statistics from January 2000 to December 2015 as case study and thereafter conducted simulation experiments on MATLAB whereby ninety five percent (95%) of data used to train the model and five percent (5%) for testing. Furthermore, the study use root mean square error (RMSE) and mean absolute percentage error (MAPE) as error metrics for model evaluation. The results show that the neuro fuzzy model have an architecture of 5:74:1 with Gaussian membership functions (2, 2, 2, 2, 2), provides RMSE of 0.44886 and MAPE 0.23384, which is far better compared to existing research studies.

Decision making both on individual and organizational level is always accompanied by the search of other’s opinion on the same. With tremendous establishment of opinion rich resources like, reviews, forum discussions, blogs, micro-blogs, Twitter etc provide a rich anthology of sentiments. This user generated content can serve as a benefaction to market if the semantic orientations are deliberated. Opinion mining and sentiment analysis are the formalization for studying and construing opinions and sentiments. The digital ecosystem has itself paved way for use of huge volume of opinionated data recorded. This paper is an attempt to review and evaluate the various techniques used for opinion and sentiment analysis.

Electrochemical impedance spectroscopy (EIS) is a key method for the characterizing the ionic and electronic conductivity of
materials. One of the requirements of this technique is a model to forecast conductivity in preliminary experiments. The aim
of this paper is to examine the prediction of conductivity by neuro-fuzzy inference with basic experimental factors such as
temperature, frequency, thickness of the film and weight percentage of salt. In order to provide the optimal sets of fuzzy
logic rule bases, the grid partition fuzzy inference method was applied. The validation of the model was tested by four
random data sets. To evaluate the validity of the model, eleven statistical features were examined. Statistical analysis of the
results clearly shows that modeling with an adaptive neuro-fuzzy is powerful enough for the prediction of conductivity.