Computational Intelligence Methods Research Articles

The Kazakh Language Requires Reform of its Writing

The article provides information about modern problems of writing the Kazakh language, the importance of its role and development in the context of mass digitization using artificial intelligence technologies and computational linguistics methods. The incorrectness of the current alphabet of the Kazakh language based on the Cyrillic alphabet is proved in connection with the inclusion of Cyrillic letters in it, denoting phonemes that are not included in its sound structure. The necessity of reforming the Kazakh writing by replacing the incorrect alphabet is substantiated. Errors and contradictions are shown in the approved version of the Kazakh alphabet based on the Latin alphabet, as well as the alphabet proposed as a replacement for the approved one, in which some previous errors are repeated. In both cases, no analysis and clarification of the sound system of the Kazakh language, which is the basis of any alphabet, is carried out. In this study, to clarify the sound system of the Kazakh language, experiments were carried out to determine the articulation and acoustic features of Kazakh sounds with the help the computer programs used for many natural languages. In the articulation analysis, special attention was paid to vowels, which give rise to various contradictions in the Kazakh letter. It is proposed to use a new classification of vowels according to four binary features, rather than the traditional classification according to three binary features. Acoustic analysis uses the method of formant analysis, which is aimed at identifying certain formants in the spectrogram. The formant is obtained using a spectrograph. Quantitatively, the formants correspond to the maxima in the speech spectrum and usually appear on spectrograms as horizontal bands. After determining the composition and classification of the sound system of the Kazakh language, two variants of the alphabet based on the Latin alphabet are proposed: the first one is based on the Turkish alphabet using diacritical marks; the second is based on the English alphabet using digraphs. The second option offers ways to solve problems that arise when using digraphs. In conclusion, information is provided on the ongoing and ongoing work in Kazakhstan related to the creation of smart systems in the Kazakh language based on the methods and technologies of artificial intelligence and computational linguistics, the results of which are reflected in the list of sources.

Magnetocaloric effect modeling of dysprosium-transition metal based intermetallic alloys for magnetic refrigeration application using hybrid genetic algorithm based support vector regression intelligent method.

Intermetallic alloy containing rare earth dysprosium ions with the associated unfilled 4f shell electrons and sub-lattice of 3d-transition metal, results into fascinating magnetic properties which are useful for green refrigeration technological application. Magnetocaloric effect remains the fundamental principle upon which magnetic refrigeration technology is based while this cooling technology has advantages of cost effectiveness, high efficiency and environmental friendliness as compared with the existing conventional gas compression systems. Maximum magnetic entropy change (which controls the hugeness of magnetocaloric effect) of intermetallic alloy Dy-T-X (where T = transition metal and X = any other metal or nonmetal) is modeled in this work using hybrid genetic algorithm based support vector regression (GSVR) computational intelligent method with applied magnetic field, ionic concentration and ionic radii descriptors. The developed GSVR-G model with kernel Gaussian function outperforms GSVR-P model with polynomial function with improvement of 85.23%, 78.82% and 78.67% on the basis of the computed correlation coefficient (CC), mean absolute error (MAE) and root mean square error (RMSE) on testing sample, respectively. The developed model further investigates the influence of applied external magnetic field on magnetocaloric effect of DyCuAl intermetallic alloy. The developed models in this work circumvent experimental challenges of magnetocaloric effect determination while the recorded precision of the developed model further opens doors for possible exploration of these intermetallic compounds for addressing environmental challenges associated with the present system of cooling.

Enhanced Solutions to Intuitionistic Fuzzy Multiobjective Linear Fractional Optimization Problems via Lexicographic Method

Optimization involving fuzzy numbers generally, and intuitionistic fuzzy numbers particularly is more and more an essential part of any computational intelligence method; and might be of interest in modeling fuzzy control systems, or carrying out a fuzzy sensitivity analysis. The recent literature includes many research papers related to both theoretical modelling, and practical implementation of fuzzy decision support systems. Fuzzy optimization is one of the main part in such decisionmaking tools. A realistic solution to a fuzzy optimization problem is always desired, and similarly, a Pareto optimal solution to a multiple objective optimization problem is always a need. In this paper we analyze the shortcomings; and eliminate the weaknesses of a solution approach from the literature proposed by Moges et al. in 2023. Firstly, Moges et al. used an accuracy function to de-fuzzify the original intuitionistic fuzzy optimization problem; then linearized the obtained crisp problem; and finally involved fuzzy goals to solve the derived multiple objective linear problem. We present two faulty key points of their approach, namely the de-fuzzification and linearization steps; prove the inappropriateness of their results; and propose some improvements. The final aim in addressing the first key point is to clarify the border between defuzzifications made in accordance to true theoretical statements, and those made for modeling reasons. The methodological improvement we propose is related to the second key point and it assures that Pareto optimal solutions - highly required in multiple objective optimization - are obtained. To illustrate our point of view, we use a numerical example from the literature, and report better numerical results derived by the improved methodology.

Fractal Geometry Meets Computational Intelligence: Future Perspectives.

Characterizations in terms of fractals are typically employed for systems with complex and multiscale descriptions. A prominent example of such systems is provided by the human brain, which can be idealized as a complex dynamical system made of many interacting subunits. The human brain can be modeled in terms of observable variables together with their spatio-temporal-functional relations. Computational intelligence is a research field bridging many nature-inspired computational methods, such as artificial neural networks, fuzzy systems, and evolutionary and swarm intelligence optimization techniques. Typical problems faced by means of computational intelligence methods include those of recognition, such as classification and prediction. Although historically conceived to operate in some vector space, such methods have been recently extended to the so-called nongeometric spaces, considering labeled graphs as the most general example of such patterns. Here, we suggest that fractal analysis and computational intelligence methods can be exploited together in neuroscience research. Fractal characterizations can be used to (i) assess scale-invariant properties and (ii) offer numeric, feature-based representations to complement the usually more complex pattern structures encountered in neurosciences. Computational intelligence methods could be used to exploit such fractal characterizations, considering also the possibility to perform data-driven analysis of nongeometric input spaces, therby overcoming the intrinsic limits related to Euclidean geometry.

A computational intelligence approach for solar photovoltaic power generation forecasting

This article describes an approach applying computational intelligence methods for the problem of forecasting solar photovoltaic power generation at country level. Precise forecast of power generation plays a vital role in designing a dependable photovoltaic power generation system. The computed predictions enable the implementation of efficient planning, management, and distribution strategies for the generated power, ultimately enhancing the performance and efficiency of the system. The study analyzes and compares artificial neural network approaches for a specific case study using real solar photovoltaic power generation data from Uruguay in the period 2018 to 2022. Several artificial neural network architectures are evaluated for forecasting. The main results indicate that the approach applying a combination of Encoder-Decoder and Long Short Term Memory artificial neural networks is the most effective method for the addressed forecasting problem. The approach yielded promising results, with an average mean error value of 0.09, improving over the other artificial neural network architectures. Even better results were obtained for sunny days. The generated forecasts hold significant value for its application in planning and scheduling processes, aiming to enhance the overall quality of service of the electricity grid.

Enhancing anomaly detection in electrical consumption profiles through computational intelligence

The advancement of society and the raising of people's standards of living depend heavily on electricity in today's world. The "zero energy buildings" idea, which recommends that buildings become self-sufficient in renewable energy to prevent the emission of CO2 into the environment, is now one of the most significant initiatives connected to building energy efficiency. This article describes a computational intelligence method to detect anomalous variations in a facility's energy use and infer a potential cause of such changes. The model is built using five sets of historical power consumption data from three buildings spread across four nations (Ecuador, Spain, France, and Canada), which are categorized based on the anomaly type each piece of data represents. Through a statistical study of the confidence interval, the proposed method, first determines the consumption patterns for each day of the week in each of the building's data sets. After normalizing the day to be studied toward its "Z" value, it is then cataloged using a machine learning model. The proposed method is evaluated in comparison to a purely statistical method called SAEEC methodology and it is discovered that the proposed method offers a relative improvement in accuracy, false positive rate (FPR), and false negative rate (FNR) of 12.41%, 42, 36%, and 42.45%, respectively, for the detection of atypical values in electrical energy consumption.

Sustainable supply chain resilience for logistics problems: Empirical validation using robust and computational intelligence methods

A supply chain's ability to avoid an unexpected disruption or quickly return to a normal situation defines supply chain resilience. Prior works have emphasized the integration of supply chain resilience and sustainable development goals (sustainable supply chain resilience), which is the focus of this paper. This paper proposes a four-level supply chain consisting of suppliers, intermodal hubs, manufacturers, and distributors along with a heterogeneous fleet of vehicles. Two types of manufacturers are included: large and small. The small manufacturers are moveable. To contribute to sustainable logistics problems, the objectives are cost, carbon emissions, and job opportunities when the appropriate supply chain members and flow of materials are optimized. To contribute to supply chain resilience, this paper presents a new robust approach and a new multiobjective algorithm to capture uncertain parameters (e.g., supply cost and employee availability). A numerical example based on empirical data shows the applicability of the proposed model.

Multi-Stage Operation Optimization of PV-Rich Low-Voltage Distribution Networks

The high expansion of a variable and intermittent nature of distributed generation, such as photovoltaics (PV), can cause technical issues in existing distribution networks (DN). In addition to producing electrical energy, PVs are inverter-based sources, and can help conventional control mechanisms in mitigating technical issues. This paper proposes a multi-stage optimal power flow (OPF)-based mixed-integer non-linear programming (MINLP) model for improving an operation state in LV PV-rich DN. A conventional control mechanism such as on load tap changer (OLTC) is used in the first stage to mitigate overvoltage caused by PVs. The second stage is related to reducing losses in DN using reactive power capabilities from PVs, which defines the optimization problem as a fully centralized observed from the distribution system operator’s (DSO) point of view. The optimization problem is realized under the co-simulation approach in which the power system analyzer and computational intelligence (CI) optimization method interact through an interface. This approach allows keeping the original MINLP model without approximations and using any computational intelligence method. OpenDSS is used as a power system analyzer, while particle swarm optimization (PSO) is used as a CI optimization method in this paper. Detailed case studies are performed and analyzed over a single-day period. To study validation and feasibility, the proposed model is evaluated on the IEEE LV European distribution feeder. The obtained results suggest that a combination of conventional control mechanisms (OLTC) and inverter-based sources (PVs) represent a promising solution for DSO and can serve as an alternative control method in active distribution networks.

Intelligent computational methods for economics

Intelligent computational methods for economics

Computational-Intelligence-Based Scheduling with Edge Computing in Cyber-Physical Production Systems.

Real-time performance and reliability are two critical indicators in cyber-physical production systems (CPPS). To meet strict requirements in terms of these indicators, it is necessary to solve complex job-shop scheduling problems (JSPs) and reserve considerable redundant resources for unexpected jobs before production. However, traditional job-shop methods are difficult to apply under dynamic conditions due to the uncertain time cost of transmission and computation. Edge computing offers an efficient solution to this issue. By deploying edge servers around the equipment, smart factories can achieve localized decisions based on computational intelligence (CI) methods offloaded from the cloud. Most works on edge computing have studied task offloading and dispatching scheduling based on CI. However, few of the existing methods can be used for behavior-level control due to the corresponding requirements for ultralow latency (10 ms) and ultrahigh reliability (99.9999% in wireless transmission), especially when unexpected computing jobs arise. Therefore, this paper proposes a dynamic resource prediction scheduling (DRPS) method based on CI to achieve real-time localized behavior-level control. The proposed DRPS method primarily focuses on the schedulability of unexpected computing jobs, and its core ideas are (1) to predict job arrival times based on a backpropagation neural network and (2) to perform real-time migration in the form of human-computer interaction based on the results of resource analysis. An experimental comparison with existing schemes shows that our DRPS method improves the acceptance ratio by 25.9% compared to the earliest deadline first scheme.

An Overview of Deep Learning Techniques for Road Detection in High-Resolution Satellite Images

One of the main and most difficult problems in computer vision research has been removing road networks from high-resolution remote sensing photos. Numerous applications, such as urban planning, traffic management, disaster response, and environmental monitoring, depend on the accurate identification of road networks. The field of artificial intelligence has witnessed tremendous advancements and advances in road extraction technology, especially in the area of deep learning. This research provides an extensive analysis of deep learning techniques for road extraction from remote sensing imagery, emphasizing how these technologies improve process accuracy and efficiency. Deep learning techniques are divided into fully supervised, semi-supervised, and unsupervised learning approaches, each having more focused subcategories, depending on the kind of annotated data that is used. The tenets, advantages, and disadvantages of these approaches are contrasted and examined. The review also provides an overview of the metrics applied to assess road extraction algorithms and the high-resolution remote sensing picture datasets used for these kinds of activities. In conclusion, we go over the main obstacles and potential applications of computational intelligence methods to enhance the precision, automation, and intelligence of road network extraction.

Exploration and quantification of magnetocaloric effect in EuTiO3 perovskite using extreme learning machine intelligent computational method

Perovskite EuTiO3 is a multi-ferroic titanate oxide material that exhibits huge magnetocaloric effect at low magnetic field without associated hysteresis loss due to the magnetic field and thermal energy. These unique features coupled with the presence of strong coupling existing between spin and lattice as well as the possibility to switch the magnetic ground state between ferromagnetism and anti-ferromagnetism, foster the potentials of this material for magnetocaloric effect based cooling technology application. Maximum magnetic entropy change of perovskite EuTiO3 determines the strength of magnetocaloric effect in the material and ultimately influences the usability of this material for cooling technology. However, experimental measurement of maximum magnetic entropy change is challenging and consumes appreciable resources and time. Maximum magnetic entropy change of intrinsic and doped EuTiO3 perovskite is modeled in this work using extreme learning machine intelligent algorithm (ELMIA) with the ionic radii, elemental concentration and applied magnetic field descriptors. The developed ELMIA-SINE with sine activation function performs better than ELMIA-SIG model with sigmoid activation function with performance superiority of 42.46%, 44.58% and 1.76% using root mean square error, mean absolute error and correlation coefficient performance yardstick, respectively. The developed ELMIA-SINE model further outperforms the existing particle swarm based support vector regression model with polynomial (PSVR-P) and Gaussian (PSVR-G) function with performance improvement of 37.03% and 14.11%, respectively. The dependence of applied magnetic field and dopant concentration on hugeness of magnetocaloric effect was established using the developed model. The precision and the ease of implementation of the developed ELMIA model as compared with the existing models would definitely facilitate exploration of magnetocaloric effect based system of refrigeration for addressing the present energy crisis.

A hybrid fuzzy logic/genetic algorithm model based on experimental data for estimation of cuttings concentration during drilling

Adequate hole cleaning is a key to economic drilling operations. However, cuttings transport during drilling is affected by different operational parameters, geometric parameters and, cuttings and drilling fluid related parameters which makes the cuttings transport phenomenon a complicated modeling exercise. This complexity has led to the use of computational intelligence (CI) methods such as fuzzy logic and genetic algorithm (GA). A total of 11 parameters are considered in the current work while developing Mamdani type fuzzy logic model using 702 experimental observations. The developed FL model (FL2) is further combined with GA to optimize fuzzy rule weights which has led to the development of a hybrid CI model termed FL3 in this paper. The performance of FL3 over two test datasets is compared with FL2 and a previously published FL model (Chowdhury and Hovda, 2022) termed FL1 (built on 509 experimental observations and 195 fuzzy rules). Model comparison using three goodness of fit metrics (R2, RMSE and MAE) shows that the hybrid CI model, FL3, outperforms both FL1 and FL2 over both test datasets. R2 value of FL3 over both test datasets is ‘very good’ following (Moriasi et al., 2007). The difference of FL3 cuttings concentration (Cc) estimates and the measured Cc values is less than 1% for more than two thirds of the experimental data for both test datasets collected at two different geographical locations by two different research groups – TUDRP and SINTEF. The reasons behind FL3 outperforming FL1 and FL2 are data randomization, modified fuzzy sets for eccentricity and drillstring rotation based on the exploratory data analysis published earlier (Chowdhury et al., 2023), use of more experimental data and the optimization of fuzzy rule weights by GA.

Predicting active compounds for lung cancer based on quantitative structure-activity relationships

<span lang="EN-US">Recently, advancements in computational and artificial intelligence (AI) methods have contributed in improving research results in the field of drug discovery. In fact, machine learning techniques have proven to be especially effective in this regard, aiding in the development of new drug variants and enabling more precise targeting of specific disease mechanisms. In this paper, we propose to use a quantitative structure-activity relationship-based approach for predicting active compounds related to non-small cell lung cancer. Our approach uses a neural network classifier that learns from sequential structures and chemical properties of molecules, as well as a gradient boosting tree classifier to conduct comparative analysis. To evaluate the contribution of each feature, we employ Shapley additive explanations (SHAP) summary plots to perform features selection. Our approach involves a dataset of active and non-active molecules collected from ChEMBL database. Our results show the effectiveness of the proposed approach when it comes to predicting accurately active compounds for lung cancer. Furthermore, our comparative analysis reveals important chemical structures that contribute to the effectiveness of the compounds. Thus, the proposed approach can greatly enhance the drug discovery pipeline and may lead to the development of new and effective treatments for lung cancer.</span>

Investigation of computational intelligence methods in forecasting at financial markets

The work considers intelligent methods for solving the problem of short- and middle-term forecasting in the financial sphere. LSTM DL networks, GMDH, and hybrid GMDH-neo-fuzzy networks were studied. Neo-fuzzy neurons were chosen as nodes of the hybrid network, which allows to reduce computational costs. The optimal network parameters were found. The synthesis of the optimal structure of hybrid networks was performed. Experimental studies of LSTM, GMDH, and hybrid GMDH-neo-fuzzy networks with optimal parameters for short- and middle-term forecasting have been conducted. The accuracy of the obtained experimental predictions is compared. The forecasting intervals for which the application of the researched artificial intelligence methods is the most expedient have been determined.

Short-Term Natural Gas and Carbon Price Forecasting Using Artificial Neural Networks

Methods of computational intelligence show a high potential for short-term price forecasting of the energy market as they offer the possibility to cope with the complexity, multi-parameter dependency, and non-linearity of pricing mechanisms. While there is a large number of publications applying neural networks to the prediction of electricity prices, the analysis of natural gas and carbon prices remains scarce. Identifying a best practice from the literature, this study presents an iterative approach to optimize both the input values and network configuration of neural networks. We apply the approach to the natural gas and carbon market, sequentially testing autoregressive and exogenous explanatory variables as well as different neural network architectures. We subsequently discuss the influence of architectural properties, input parameters, data preparation, and the models’ resilience to singular events. Results show that the selection of appropriate lags of gas and carbon prices to account for autoregressive properties of the respective time series leads to a high degree of forecasting accuracy. Additionally, including ambient temperature data can slightly reduce errors of natural gas price forecasting whereas carbon price predictions benefit from electricity prices as a further explanatory input. The best configurations presented in this contribution achieve a root mean square error (RMSE) of 0.64 EUR/MWh (natural gas prices) corresponding to a normalized RMSE of 0.037 and 0.33 EUR/tCO2 (carbon prices) corresponding to a normalized RMSE of 0.023.

Bio-inspired swarm-intelligent with machine learning framework for prediction and classification of lung cancer

Early detection of lung cancer nodules, which is often dependent on a tomography scan filmic examination, greatly increases the probability of survival. Earlier tumor detection reduces lung cancer mortality by increasing the likelihood of successful therapy. Traditionally, radiologists have analyzed medical pictures for signs of lung cancer using a laborious and inaccurate systematic approach. The confidentiality and integrity of medical data have become a serious challenge for healthcare applications as a result of the tremendous improvement in the transmission of medical data in the healthcare sector. This research makes use of computational intelligence methods. In this research, a brand-new Enhanced vortex search algorithm optimized Support Vector Machine (EVSAO-SVM) is developed for detection and classification. Preprocessing using the Gaussian Filter, segmentation with Otsu thresholding, feature extraction with local binary patterns (LPB), and classification and prediction with the EVSAO-SVM are the processes that are simulated. Regarding performance criteria including accuracy, sensitivity, specificity, and F1-measure, this study indicates the suggested framework's superiority over the traditional approaches. The results of the tests conducted demonstrate that the suggested model can reach up to 95.42 percent sensitivity, 96.24 percent accuracy, 98.92 percent specificity, and 94.26 percent F1 measure.

SenticGAT: Sentiment Knowledge Enhanced Graph Attention Network for Multi-view Feature Representation in Aspect-based Sentiment Analysis

Currently, computational intelligence methods, especially artificial neural networks, are increasingly applied to many scenarios. We mainly attempt to explore the task of fine-grained sentiment classification of review data through computational intelligence methods, especially artificial neural networks, and this task is also known as aspect-based sentiment analysis (ABSA). We propose a new technique called SenticGAT which is a multi-view features fusion model enhanced by an external sentiment database. We encode the external sentiment information into the syntactic dependency tree to obtain an enhanced graph with rich sentiment representation. Then we obtain multi-view features including semantics, syntactic, and sentiment features through GAT based on the enhanced graph by external knowledge. We also design a new strategy for fusing multi-view features using the feature parallel frame and convolution method. Eventually, the sentiment polarity of a specific aspect is determined based on the completely fused multi-view features. Experimental results on four public benchmark datasets demonstrate that our method is effective and sound. And it performs superiorly to existing approaches in fusion multiple-view features.

Microgrid planning based on computational intelligence methods for rural communities: A case study in the José Painecura Mapuche community, Chile

Microgrids (MGs) are sustainable solutions for rural zone electrification that use local renewable resources. However, only careful planning at the start of an MG project can ensure its future optimal operation. In this paper, a novel methodology for MG planning by using the uncertainty characterization of renewable resources and demand is presented. Additionally, a model of electricity consumption is proposed and applied in an isolated rural community. In such communities, consumption patterns typically need to be derived as model inputs because consumption measurements are not available for the planning stage. To obtain these inputs, clustering algorithms based on self-organizing maps (SOMs) and fuzzy c-means are used to classify the families of the community given sociodemographic information obtained via surveys. Subsequently, Markov chains (MCs) are employed to generate consumption patterns based on consumption measurements in some dwellings and surveys applied to the community. The nonlinearities and uncertainties associated with renewable resources and consumption are modeled by using prediction interval (PI) models. These PI models provide the required consumption and generation scenarios for deriving the optimal sizing and topological information to address the MG planning problem. The results of the robust planning approach based on scenarios are useful at the feasibility and design phases of an MG project. The proposed methodology is successfully applied to MG planning for a rural Mapuche community, where a conservative criterion was considered to minimize the investment risk. This criterion corresponds to the worst-case scenario in which the demand increases by 19.9% compared to that of the baseline scenario and a lower energy cost is obtained. However, the net present cost and operational costs increase by 14% and 11.75% compared to those of the baseline scenario, respectively.

Riverside Landslide Susceptibility Overview: Leveraging Artificial Neural Networks and Machine Learning in Accordance with the United Nations (UN) Sustainable Development Goals

Riverside landslides present a significant geohazard globally, posing threats to infrastructure and human lives. In line with the United Nations’ Sustainable Development Goals (SDGs), which aim to address global challenges, professionals in the field have developed diverse methodologies to analyze, assess, and predict the occurrence of landslides, including quantitative, qualitative, and semi-quantitative approaches. With the advent of computer programs, quantitative techniques have gained prominence, with computational intelligence and knowledge-based methods like artificial neural networks (ANNs) achieving remarkable success in landslide susceptibility assessments. This article offers a comprehensive review of the literature concerning the utilization of ANNs for landslide susceptibility assessment, focusing specifically on riverside areas, in alignment with the SDGs. Through a systematic search and analysis of various references, it has become evident that ANNs have emerged as the preferred method for these assessments, surpassing traditional approaches. The application of ANNs aligns with the SDGs, particularly Goal 11: Sustainable Cities and Communities, which emphasizes the importance of inclusive, safe, resilient, and sustainable urban environments. By effectively assessing riverside landslide susceptibility using ANNs, communities can better manage risks and enhance the resilience of cities and communities to geohazards. While the number of ANN-based studies in landslide susceptibility modeling has grown in recent years, the overarching objective remains consistent: researchers strive to develop more accurate and detailed procedures. By leveraging the power of ANNs and incorporating relevant SDGs, this survey focuses on the most commonly employed neural network methods for riverside landslide susceptibility mapping, contributing to the overall SDG agenda of promoting sustainable development, resilience, and disaster risk reduction. Through the integration of ANNs in riverside landslide susceptibility assessments, in line with the SDGs, this review aims to advance our knowledge and understanding of this field. By providing insights into the effectiveness of ANNs and their alignment with the SDGs, this research contributes to the development of improved risk management strategies, sustainable urban planning, and resilient communities in the face of riverside landslides.