Machine Intelligent Systems Research Articles

A Critical Review on Selected Fuzzy Min-Max Neural Networks and Their Significance and Challenges in Pattern Classification

At present, pattern classification is one of the most important aspects of establishing machine intelligence systems for tackling decision-making processes. The fuzzy min–max (FMM) neural network combines the operations of an artificial neural network and fuzzy set theory into a common framework. FMM is considered one of the most useful neural networks for pattern classification. This paper aims to 1) analyze the FMM neural network in terms of its impact in addressing pattern classification problems; 2) examine models that are proposed based on the original FMM model (i.e., existing FMM-based variants); 3) identify the challenges associated with FMM and its variants, and; 4) discuss future trends and make recommendations for improvement. The review is conducted based on a methodical protocol. Through a rigorous searching and filtering process, the relevant studies are extracted and comprehensively analyzed to adequately address the defined research questions. The findings indicate that FMM plays a critical role in providing solutions to pattern classification issues. The FMM model and a number of FMM-based variants are identified and systematically analyzed with respect to their aims, improvements introduced and results achieved. In addition, FMM and its variants are critically analyzed with respect to their benefits and limitations. This paper shows that the existing FMM-based variants still encounter issues in terms of the learning process (expansion, overlap test, and contraction), which influence the classification performance. Based on the review findings, research opportunities are suggested to propose a new model to enhance the number of existing FMM models, particularly in terms of their learning process by minimizing hyperbox overlap pertaining to different classes as well as avoiding membership ambiguity of the overlapped region. In short, this review provides a comprehensive and critical reference for researchers and practitioners to leverage FMM and its variants for undertaking pattern classification tasks.

Adaptive multi-parent crossover GA for feature optimization in epileptic seizure identification

EEG signal analysis involves multi-frequency non-stationary brain waves from multiple channels. Segmenting these signals, extracting features to obtain the important properties of the signal and classification are key aspects of detecting epileptic seizures. Despite the introduction of several techniques, it is very challenging when multiple EEG channels are involved. When many channels exist, a spatial filter is required to eliminate noise and extract relevant information. This adds a new dimension of complexity to the frequency feature space. In order to stabilize the classifier of the channels, feature selection is very important. Furthermore, and to improve the performance of a classifier, more data is required from EEG channels for complex problems. The increase of such data poses some challenges as it becomes difficult to identify the subject dependent bands when the channels increase. Hence, an automated process is required for such identification.The proposed approach in this work tends to tackle the multiple EEG channels problem by segmenting the EEG signals in the frequency domain based on changing spikes rather than the traditional time based windowing approach. While to reduce the overall dimensionality and preserve the class-dependent features an optimization approach is used. This process of selecting an optimal feature subset is an optimization problem. Thus, we propose an adaptive multi-parent crossover Genetic Algorithm (GA) for optimizing the features used in classifying epileptic seizures. The GA-based approach is used to optimize the various features obtained. It encodes the temporal and spatial filter estimates and optimize the feature selection with respect to the classification error. The classification was done using a Support Vector Machine (SVM).The proposed technique was evaluated using the publicly available epileptic seizure data from the machine learning repository of the UCI center for machine learning and intelligent systems. The proposed approach outperforms other ones and achieved a high level of accuracy. These results, indicate the ability of a multi-parent crossover GA in optimizing the feature selection process in EEG classification.

A local random walk model for complex networks based on discriminative feature combinations

Networks have become increasingly important to model many complex systems. This powerful representation has been employed in different tasks of artificial intelligence including machine learning, expert and intelligent systems. Link prediction, a branch of network pattern recognition, is the most fundamental and essential problem for complex network analysis. However, most existing link-prediction methods only consider a network’s topology structures, and in doing so, these methods miss the opportunity to use nodes’ attribute information. We present a combined approach here that uses nodes’ attribute information and topology structure to direct link prediction. First, we propose a discriminative feature combinations selection method. Specifically, we present a novel mathematics inference to detail discriminative feature combinations. Second, based on the selected feature combinations, we aggregate the network, and further compute each feature combination’s contributing degree to the link’s formation, called the strength of feature combination. Third, we apply discriminative feature combinations into a local random walk model; in particular, we compute and redistribute the random walk particle’s transfer probability in terms of each feature combination’s strength, which makes the transfer probability depend on feature combinations satisfied by each node’s edges. Finally, we predict links in complex networks based on the improved random walk model. Experimental results on real-life complex network datasets demonstrate that, compared to other baseline methods, using discriminative feature combinations and topology structures in tandem strengthens prediction performance remarkably.

Labor management and social and labor relations in the digital economy

Social partnership is the most important type of social and labor relations in the modern economy, which is based on the active participation of trade unions in the enterprises management, payroll rate determination, and improvement of its conditions. In addition, in the framework of the conditions of extensive application of digital technologies the employees being outside the premises of the employer’s organization began to perform a significant part of the functions of many organizations, and for a certain category of employees, the possibility of their replacement with intelligent machine systems will become the reality in the near future. The paper shows that one of the optimality criteria for the parameters of the social system is the country’s living standards. The proposed criterion takes into account GDP per capita, the Gini ratio and the crime rate. The paper also shows that in the digital economy one of the main factors of the living standards growth is the application of new scientific results. The specific measures are proposed to create new forms of social and labor relations in the digital economy.

Kinect Sensor Gesture and Activity Recognition: New Applications for Consumer Cognitive Systems

Cognitive consumer electronics (CE), the fastest-growing sector worldwide that is driven by machine intelligence and cognitive systems, is triggered and enabled by audio- and video-capturing devices, smart sensors, health- and fitness-monitoring devices, security and education electronics, and intelligent systems. Smart consumer sensors and cognitive systems are synergized through the Internet of Things (IoT) for optimal information sharing, communication, real-time updates, data analytics, and enhanced support for decision making. Biometric-based devices, originally intended for large-scale applications in airports, border controls, disaster zones, or refugee migration zones, are enabling a wide range of applications in commercial and consumer sectors as standalone systems or with interconnected sensor networks. This article introduces the applications of Microsoft Kinect in cognitive systems for smart CE, and, using Kinect sensors, a human-behavior cognition technology is presented for gesture and activity recognition. As a novel front end of pervasive cognitive systems, the challenges and applications of a Kinect sensor-based system will be explored in CE, such as smart automobiles, health care, surveillance, and activity recognition.

Engineering Courses on Computational Thinking Through Solving Problems in Artificial Intelligence

Computational thinking sits at the core of every engineering and computing related discipline. It has increasingly emerged as its own subject in all levels of education. It is a powerful cornerstone for cognitive development, creative problem solving, algorithmic thinking and designs, and programming. How to effectively teach computational thinking skills poses real challenges and creates opportunities. Targeting entering computer science and engineering undergraduates, we resourcefully integrate elements from artificial intelligence (AI) into introductory computing courses. In addition to comprehension of the essence of computational thinking, practical exercises in AI enable inspirations of collaborative problem solving beyond abstraction, logical reasoning, critical and analytical thinking. Problems in machine intelligence systems intrinsically connect students to algorithmic oriented computing and essential mathematical foundations. Beyond knowledge representation, AI fosters a gentle introduction to data structures and algorithms. Focused on engaging mental tool, a computer is never a necessity. Neither coding nor programming is ever required. Instead, students enjoy constructivist classrooms designed to always be active, flexible, and highly dynamic. Learning to learn and reflecting on cognitive experiences, they rigorously construct knowledge from collectively solving exciting puzzles, competing in strategic games, and participating in intellectual discussions.

Hybrid robust iris recognition approach using iris image pre‐processing, two‐dimensional gabor features and multi‐layer perceptron neural network/PSO

Computational intelligence is employed to solve factual and complicated global problems, though neural networks (NNs) and evolutionary computing have also affected these issues. Biometric traits are applicable for detecting crime in security systems because they offer attractive features such as stability and uniqueness. Although various methods have been proposed for this objective, feature shortcomings such as computational complexity, long run times, and high memory consumption remain. The current study proposes a novel human iris recognition approach based on a multi-layer perceptron NN and particle swarm optimisation (PSO) algorithms to train the network in order to increase generalisation performance. A combination of these algorithms was used as a classifier. A pre-processing step was performed on the iris images to improve the results and two-dimensional gabor kernel feature extraction was applied. The data was normalised, trained, and tested using the proposed method. A PSO algorithm was applied to train the NN for data classification. The experimental results show that the proposed method performs better than many other well-known techniques. The benchmark Chinese Academy of Science and Institute of Automation (CASIA)-iris V3 and Center for Machine Learning and Intelligent Systems at the University of California, Irvine (UCI) machine learning repository datasets were used for testing and comparison.

Noise-assisted multivariate empirical mode decomposition for multichannel EMG signals

BackgroundEnsemble Empirical Mode Decomposition (EEMD) has been popularised for single-channel Electromyography (EMG) signal processing as it can effectively extract the temporal information of the EMG time series. However, few papers examine the temporal and spatial characteristics across multiple muscle groups in relation to multichannel EMG signals.ExperimentThe experimental data was obtained from the Center for Machine Learning and Intelligent Systems, University of California Irvine (UCI). The data was donated by the Nueva Granada Military University and the Technopark node Manizales in Colombia. The databases of 11 male subjects from the healthy group were taken into the study. The subjects undergo three exercise programs, leg extension from a sitting position (sitting), flexion of the leg up (standing), and gait (walking), while four electrodes were placed on biceps femoris (BF), vastus medialis (VM), rectus femoris (RF), and semitendinosus (ST).MethodsBased on the experimental data, a comparative study is provided by assessing the Empirical Mode Decomposition (EMD)-based approaches, EEMD, Multivariate EMD (MEMD), and Noise-Assisted MEMD (NA-MEMD). The outcomes from these approaches are then quantitatively estimated on the basis of three criterions, the number of Intrinsic Mode Functions (IMFs), mode-alignment and mode-mixing.ResultsBoth MEMD and NA-MEMD methods (except EEMD) can guarantee equal numbers of IMFs. For mode-alignment and mode-mixing, NA-MEMD is optimal compared with MEMD and EEMD, and MEMD is merely better than EEMD.ConclusionsThis study proposes the NA-MEMD approach for multichannel EMG signal processing. This finding implies that NA-MEMD is effective for simultaneously analysing IMFs based frequency bands. It has a vital clinical implication in exploring the neuromuscular patterns that enable the multiple muscle groups to coordinate while performing the functional activities of daily living.

Subjectivity of political knowledge: gonoseological dimension

The article explores the specifics of the separation of political gnoseology, political epistemology and political cognition as directions of penetration of the subject of knowledge into the essence of political phenomena and processes at the methodological level. It is emphasized that the classical and non-classical theory of cognition has paved the way for a modern scientific understanding of experimental methodology and generally accepted standards of scientific novelty. The attention is focused on the formation of one-dimensional political views on political cognition as a reflective reflection of the political world in a certain period of time by the actual means of cognition. It is emphasized that the question of the significance of the classical directions of cognition, mono-ideological and monodoctrynnyh structures of political world perception is actualized. It is argued that in the conditions of informatization and the growth of the importance of machine intelligence systems in the practice of obtaining new knowledge, it is necessary to find out the problem of the value of artificial intelligence and non-human systems of thinking for political knowledge of the future. It is substantiated that the scientific paradigm of thinking defines standard requirements for the effectiveness and efficiency of political activity. Considered the importance of part of political knowledge, which may not meet the criteria of natural or technical scientific.

Parallel driving in CPSS: a unified approach for transport automation and vehicle intelligence

The emerging development of connected and automated vehicles imposes a significant challenge on current vehicle control and transportation systems. This paper proposes a novel unified approach, Parallel Driving, a cloud-based cyberphysical-social systems U+0028 CPSS U+0029 framework aiming at synergizing connected automated driving. This study first introduces the CPSS and ACP-based intelligent machine systems. Then the parallel driving is proposed in the cyber-physical-social space, considering interactions among vehicles, human drivers, and information. Within the framework, parallel testing, parallel learning and parallel reinforcement learning are developed and concisely reviewed. Development on intelligent horizon U+0028 iHorizon U+0028 and its applications are also presented towards parallel horizon. The proposed parallel driving offers an ample solution for achieving a smooth, safe and efficient cooperation among connected automated vehicles with different levels of automation in future road transportation systems.

Path planning for support heads in mirror-milling machining system

To reduce dynamic deflection and suppress vibration during the end-milling process of large thin-plate workpiece, an intelligent mirror-milling machining system is established in this paper. This system includes two movable support heads, and both of them are periodically repositioned which affix themselves to the workpiece by the way of vacuum adsorption in order to improve stability of the workpiece during the milling process. One critical point of this task is how to ensure support heads move reasonably and simultaneously and support to the workpiece stably. Therefore, this paper developed a practical method for the planning of support head moving path (SHMP) under a given cutting path. This method involves the finite difference method (FDM), artificial neural network (ANN), and genetic algorithm (GA). In particular, dynamic vibration response of four-edge clamped thin-plate workpiece is calculated with FDM; to reduce computing time, ANN is applied to establish the relation between support head location and workpiece dynamic vibration amplitude; and as SHMP is a multi-objective optimization problem, GA is implemented with the classical weighted sum method to find a solution. Finally, this method is verified and simulated under two specific cases; the final paths obviously show that there will always be at least one support head support workpiece during the machining process.

Shared and service-oriented CNC machining system for intelligent manufacturing process

To improve efficiency, reduce cost, ensure quality effectively, researchers on CNC machining have focused on virtual machine tool, cloud manufacturing, wireless manufacturing. However, low level of information shared among different systems is a common disadvantage. In this paper, a machining database with data evaluation module is set up to ensure integrity and update. An online monitoring system based on internet of things and multi-sensors “feel” a variety of signal features to “percept” the state in CNC machining process. A high efficiency and green machining parameters optimization system “execute” service-oriented manufacturing, intelligent manufacturing and green manufacturing. The intelligent CNC machining system is applied in production. CNC machining database effectively shares and manages process data among different systems. The prediction accuracy of online monitoring system is up to 98.8% by acquiring acceleration and noise in real time. High efficiency and green machining parameters optimization system optimizes the original processing parameters, and the calculation indicates that optimized processing parameters not only improve production efficiency, but also reduce carbon emissions. The application proves that the shared and service-oriented CNC machining system is reliable and effective. This research presents a shared and service-oriented CNC machining system for intelligent manufacturing process.

Machine Learning of Cutting Conditions in Drilling Using Artificial Neural Network

Machining has traditionally been one of the major operations within most manufacturing systems and intelligent machining will play an important role in feature manufacturing systems. This paper concerns the machine learning, specifically classification and recognition of cutting conditions in drilling process. Awareness of the cutting conditions can enhance the auto-diagnosis of an intelligent machining system. In this paper, features representing the drilling process are generated from the converted forms of thrust force and torque and extracted with wavelet packet transform (WPT) and then selected by principal component analysis (PCA). Data instances are generated from experiments with different cutting conditions including workpiece material, drill diameter, feed rate, and spindle speed. A feed-forward network trained with back-propagation method (BPNN) is applied to distinguish between the patterns of each cutting condition. The different contributions of features and the recognition results of cutting conditions are discussed.

Regional Development Classification Model using Decision Tree Approach

Regional development classification is one way to look at differences in levels of development outcomes. Some frequently used methods are the shift share, Gain index, the Iindex Williamson and Klassen typology. The development of science in the field of data mining, offers a new way for regional development data classification. This study discusses how the decision tree is used to classify the level of development based on indicators of regional gross domestic product (GDP). GDP Data Central Java and Banten used in this study. Before the data is entered into the decision tree forming algorithm, both the provincial GDP data are classified using Klassen typology. Three decision tree algorithms, namely J48, NBTRee and REPTree tested in this study using cross-validation evaluation, then selected one of the best performing algorithms. The results show that the J48 has a better accuracy rate which is equal to 85.18% compared to the algorithm NBTRee and REPTree. Testing the model is done to the six districts / municipalities in the province of Banten, and shows that there are two districts / cities are still at the development of the status quadrant relatively underdeveloped regions, namely Kota Tangerang and Kabupaten Tangerang. As for the Central Java Province, Kendal, Magelang, Pemalang, Rembang, Semarang and Wonosobo are an area with a quadrant of development also on the status of the region is relatively underdeveloped. Classification model that has been developed is able to classify the level of development fast and easy to enter data directly into the decision tree is formed. This study can be used as an alternative decision support for policy makers in order to determine the future direction of development. General Terms Machine intelligence, data mining and decision support systems.

Classification of Breast Cancer using Gini Index based Fuzzy Supervised Learning in Quest Decision Tree Algorithm

Decision tree is a dominating method of pattern classification. These trees amongst the machine learning techniques have an aptitude to handle and construe logical rules of classification. The general Decision trees face the problem on deciding boundaries in classification. A fuzzy supervised learning in Quest (SLIQ) decision tree (FS-DT) algorithm aimed at constructing a fuzzy decision boundary instead of puny decision boundaries. The intended work deals with a fuzzy supervised learning in Quest (SLIQ) decision tree (FS-DT) algorithm by calculating the Gini index at the points where the class information changes. This algorithm helps in predicting benign and malignant breast cancer cases more effectively. The breast cancer mammographic mass dataset (BI-RADS) was taken from UCI Machine Learning Repository, center for machine learning and intelligent systems. A 3-fold cross validation of train data and test data on BI-RADS dataset was used and the proposed Fuzzy SLIQ Decision Tree algorithm was applied on it. The proposed method‟s performance was superior to earlier techniques. The examined results in partitioning the benign and malignant cases using Fuzzy SLIQ Decision Tree is more promising with a classification accuracy of 81.4% which is more prominent then many of the existing classifier techniques which used BI-RADS dataset in classification of breast cancer cases.

A Sugeno-Type Fuzzy Expert System for Rough Turning

This work describes a fuzzy expert system for rough turning. In order to automate unmanned turning, safety of the process must be ensured. In addition, any quality requirements should be fulfilled and, within these constraints, productivity maximized. The traditional approach in adaptive control of machining is to keep a measured quantity, such as power, within acceptable limits. However, there have been some studies measuring distinct phenomena in machining and identifying “cutting states” based on the phenomena. By identifying cutting states corresponding to phenomena monitored by human experts, it is possible to construct an intelligent machining system emulating the decision making of a human expert. This paper concentrates on defining the requirements for the inference part of such of an intelligent machining system. This work concentrates on both functional requirements, such as capability to take into account specific cutting states. The existence of process monitoring subsystems which detect and measure the cutting phenomena is assumed. As a result, a Sugeno-type fuzzy control is suggested, and feasibility and the level of completeness of such a system are discussed and issues requiring further study are identified.

Associative memories based on cellular automata: an application to pattern recognition

This paper presents the proposal for an associative memory implemented in cellular automata for pattern recognition. Both in their learning phase and recovery, the proposed model is based on dilation and erosion operators from mathematical morphology. The model was applied to the iris plant, abalone and hepatitis databases taken from the repertoire of available databases for Machine Learning Center and Intelligent Systems at the University of California Irvine.

Rough Set Approach in Machine Learning: A Review

The Rough Set (RS) theory can be considered as a tool to reduce the input dimensionality and to deal with vagueness and uncertainty in datasets. Over the years, there has been a rapid growth in interest in rough set theory and its applications in artificial intelligence and cognitive sciences, especially in research areas such as machine learning, intelligent systems, inductive reasoning, pattern recognition, data preprocessing, knowledge discovery, decision analysis, and expert systems. This paper discusses the basic concepts of rough set theory and point out some rough set-based research directions and applications. The discussion also includes a review of rough set theory in various machine learning techniques like clustering, feature selection and rule induction. General Terms Information and decision systems, (in)discernibility , approximation spaces, rough sets, rough membership functions, reducts, decision rules, dependencies of attributes. Clustering, Rule Induction, Feature Selection

A unified granular hybrid soft computing framework for vision engineering

We propose a novel unifying framework for building a novel granular modular architecture for a machine visual system that accommodates a large spectrum of potential vision problems. Thus removing the ad hoc nature of present solutions and providing the basis for new generation of machine visual systems. Such a framework works by integrating some useful concepts from the human vision processes and adding some interesting granular functionalities of human cognition and it advocates further hybridisation of non-linear digital filters and soft computing in implementing such machine intelligent visual systems. Our focus herein will be on the low level and mid-level stages of such a framework. The goal is to build an automatic system that can be used for degraded multi-modal image processing, including x-rays, MRI, Sonar, etc. for diagnosis, recognition, registration and information fusion purposes. For illustration purposes, an investigation concerning its application to a real world problem is also provided. We are interested by an application to automatic detection and classification of patients| spines affected by idiopathic scoliosis from X-rays images.

Gaussian kernel based fuzzy rough sets: Model, uncertainty measures and applications

Kernel methods and rough sets are two general pursuits in the domain of machine learning and intelligent systems. Kernel methods map data into a higher dimensional feature space, where the resulting structure of the classification task is linearly separable; while rough sets granulate the universe with the use of relations and employ the induced knowledge granules to approximate arbitrary concepts existing in the problem at hand. Although it seems there is no connection between these two methodologies, both kernel methods and rough sets explicitly or implicitly dwell on relation matrices to represent the structure of sample information. Based on this observation, we combine these methodologies by incorporating Gaussian kernel with fuzzy rough sets and propose a Gaussian kernel approximation based fuzzy rough set model. Fuzzy T-equivalence relations constitute the fundamentals of most fuzzy rough set models. It is proven that fuzzy relations with Gaussian kernel are reflexive, symmetric and transitive. Gaussian kernels are introduced to acquire fuzzy relations between samples described by fuzzy or numeric attributes in order to carry out fuzzy rough data analysis. Moreover, we discuss information entropy to evaluate the kernel matrix and calculate the uncertainty of the approximation. Several functions are constructed for evaluating the significance of features based on kernel approximation and fuzzy entropy. Algorithms for feature ranking and reduction based on the proposed functions are designed. Results of experimental analysis are included to quantify the effectiveness of the proposed methods.