Search Results (71)

The complex and specialized terminology of financial language in Portuguese-speaking markets create significant challenges for natural language processing (NLP) applications, which must capture nuanced linguistic and contextual information to support accurate analysis and decision-making. This paper presents DeB3RTa, a transformer-based model specifically developed through a mixed-domain pretraining strategy that combines extensive corpora from finance, politics, business management, and accounting to enable a nuanced understanding of financial language. DeB3RTa was evaluated against prominent models—including BERTimbau, XLM-RoBERTa, SEC-BERT, BusinessBERT, and GPT-based variants—and consistently achieved significant gains across key financial NLP benchmarks. To maximize adaptability and accuracy, DeB3RTa integrates advanced fine-tuning techniques such as layer reinitialization, mixout regularization, stochastic weight averaging, and layer-wise learning rate decay, which together enhance its performance across varied and high-stakes NLP tasks. These findings underscore the efficacy of mixed-domain pretraining in building high-performance language models for specialized applications. With its robust performance in complex analytical and classification tasks, DeB3RTa offers a powerful tool for advancing NLP in the financial sector and supporting nuanced language processing needs in Portuguese-speaking contexts. Full article

Automatic speech recognition (ASR) and natural language processing (NLP) play key roles in advancing human–technology interactions, particularly in healthcare communications. This study aims to enhance French-language online mental health platforms through the adaptation of the QuartzNet 15 × 5 ASR model, selected for its robust performance across a variety of French accents as demonstrated on the Mozilla Common Voice dataset. The adaptation process involved tailoring the ASR model to accommodate various French dialects and idiomatic expressions, and integrating it with an NLP system to refine user interactions. The adapted QuartzNet 15 × 5 model achieved a baseline word error rate (WER) of 14%, and the accompanying NLP system displayed weighted averages of 64.24% in precision, 63.64% in recall, and an F1-score of 62.75%. Notably, critical functionalities such as ‘Prendre Rdv’ (schedule appointment) achieved precision, recall, and F1-scores above 90%. These improvements substantially enhance the functionality and management of user interactions on French-language digital therapy platforms, indicating that continuous adaptation and enhancement of these technologies are beneficial for improving digital mental health interventions, with a focus on linguistic accuracy and user satisfaction. Full article

Offshore wind-to-hydrogen production is an effective means of solving the problems of large-scale grid-connected consumption and high power transmission costs of offshore wind power. Site selection is a core component in planning offshore wind-to-hydrogen facilities, involving careful consideration of multiple factors, and is a classic multi-criteria decision-making problem. Therefore, this study proposes a multi-criteria decision-making method based on the two-dimensional linguistic cloud model to optimize site selection for offshore wind-to-hydrogen bases. Firstly, the alternative schemes are evaluated using two-dimensional linguistic information, and a new model for transforming two-dimensional linguistic information into a normal cloud is constructed. Then, the cloud area overlap degree is defined to calculate the interaction factor between decision-makers, and a multi-objective programming model based on maximum deviation-minimum correlation is established. Following this, the Pareto solution of criteria weights is solved using the non-dominated sorting genetic algorithm II, and the alternatives are sorted and selected through the cloud-weighted average operator. Finally, an index system was constructed in terms of resource conditions, planning conditions, external conditions, and other dimensions, and a case study was conducted using the location of offshore wind-to-hydrogen production bases in Shanghai. The method proposed in this study demonstrates strong robustness and can provide a basis for these multi-criteria decision-making problems with solid qualitative characteristics. Full article

Fake news detection (FND) remains a challenge due to its vast and varied sources, especially on social media platforms. While numerous attempts have been made by academia and the industry to develop fake news detection systems, research on Arabic content remains limited. This study investigates transformer-based language models for Arabic FND. While transformer-based models have shown promising performance in various natural language processing tasks, they often struggle with tasks involving complex linguistic patterns and cultural contexts, resulting in unreliable performance and misclassification problems. To overcome these challenges, we investigated an ensemble of transformer-based models. We experimented with five Arabic transformer models: AraBERT, MARBERT, AraELECTRA, AraGPT2, and ARBERT. Various ensemble approaches, including a weighted-average ensemble, hard voting, and soft voting, were evaluated to determine the most effective techniques for boosting learning models and improving prediction accuracies. The results of this study demonstrate the effectiveness of ensemble models in significantly boosting the baseline model performance. An important finding is that ensemble models achieved excellent performance on the Arabic Multisource Fake News Detection (AMFND) dataset, reaching an F1 score of 94% using weighted averages. Moreover, changing the number of models in the ensemble has a slight effect on the performance. These key findings contribute to the advancement of fake news detection in Arabic, offering valuable insights for both academia and the industry Full article

Chinese family farms are continuously expanding; they are also facing various business risks that lead to a shorter lifespan. This paper constructed a family farm business risk assessment model that combined a hesitant fuzzy linguistic term sets (HFLTS) model with a hesitant fuzzy weighted average (HFWA) operator. On the basis of the factor analysis, this study built a family farm indicator system that included the natural, technical, market, policy, society, and management risk. The HFLTS was used for the assessment of weights in pairwise comparison matrices, and the HFWA operator was used as an aggregation operator to calculate the business risk score of family farms. For our case study, a method comparison analysis was also performed to check the validity of the results obtained by our risk assessment model. Full article

The 360 degree feedback evaluation method is a multidimensional, comprehensive assessment method. Evaluators may hesitate among multiple evaluation values and be simultaneously constrained by the biases and cognitive errors of the evaluators, evaluation results are prone to unfairness and conflicts. To overcome these issues, this paper proposes a consensus-based 360 degree feedback evaluation method with linguistic distribution assessments. Firstly, evaluators provide evaluation information in the form of linguistic distribution. Secondly, utilizing an enhanced ordered weighted averaging (OWA) operator, the model aggregates multi-source evaluation information to handle biased evaluation information effectively. Subsequently, a consensus-reaching process is established to coordinate conflicting viewpoints among the evaluators, and a feedback adjustment mechanism is designed to guide evaluators in refining their evaluation information, facilitating the attainment of a unanimous evaluation outcome. Finally, the improved 360 degree feedback evaluation method was applied to the performance evaluation of the project leaders in company J, thereby validating the effectiveness and rationality of the method. Full article

Against the background of a major change in the world unseen in a century, emergencies with high complexity and uncertainty have had serious impacts on economic security and sustainable social development, making emergency management an important issue that needs to be urgently resolved, and the quality assessment of emergency information is a key link in emergency management. To effectively deal with the uncertainty of emergency information quality assessment, a new fuzzy multi-attribute assessment method is proposed in this paper. First, we propose the linguistic complex T-spherical fuzzy set (LCT-SFS), which can deal with two-dimensional problems and cope with situations in which assessment experts cannot give quantitative assessments. Then, the advanced linguistic complex T-spherical fuzzy Dombi-weighted power-partitioned Heronian mean (ALCT-SFDWPPHM) operator, which incorporates the flexibility of Dombi operations, is proposed. The partitioned Heronian mean (PHM) operator can consider attribute partitioning and attribute correlation, the power average (PA) operator can eliminate the effect of evaluation singularities, and the advanced operator can circumvent the problem of consistent or indistinguishable aggregation results, which provides a strong comprehensive advantage in the evaluating information aggregation. Finally, a fuzzy multi-attribute assessment model is constructed by combining the proposed operator with the WASPAS method and applied to the problem of assessing the quality and sensitivity of emergency information; qualitative and quantitative comparison analyses are carried out. The results show the method proposed in this paper has strong feasibility and validity and can represent uncertainty assessment more flexibly while providing reasonable and reliable results. The method can provide new ideas and methods for the quality assessment of emergency information, and promoting sustainable, efficient, and high-quality development of emergency management. Full article

Due to the complexity and uncertainty of decision-making, probabilistic linguistic term sets (PLTSs) are currently important tools for qualitative evaluation of decision-makers. The asymmetry of evaluation information can easily lead to the loss of subjective preference information for decision-makers, and the existing operation of decision-maker evaluation information fusion operators is difficult to solve this problem. To solve such problems, this paper proposes some new operational methods for PLTSs based on Dombi T-conorm and T-norm. Considering the interrelationships between the input independent variables of PLTSs, the probabilistic linguistic weighted Dombi Bonferroni mean Power average (PLWDBMPA) operators are extended and the properties of these aggregation operators are proposed. Secondly, the PLWDBMPA operator is used to fuse the evaluation information of decision-makers, avoiding the loss of decision information as much as possible. This paper uses social media platforms and web crawler technology to obtain online comments from users on decision-making to obtain the public’s attitude towards decision events. TF-IDF and Word2Vec are used to calculate the weight of alternatives on each attribute. Under traditional group decision-making methods and integrating the wisdom of the public, a novel multi-attribute group decision-making method based on TODIM method is proposed. Finally, the case study of Turkey earthquake shelter selection proves this method is scientific and effective. Meanwhile, the superiority of this method was further verified through comparisons with the PL-TOPSIS, PLWA, SPOTIS and PROMETHEE method. Full article

A well-researched field is the development of Computer Aided Diagnosis (CADx) Systems for the benign-malignant classification of abnormalities detected by mammography. Due to the nature of the breast parenchyma, there are significant uncertainties about the shape and geometry of the abnormalities that may lead to an inaccurate diagnosis. These same uncertainties give mammograms a fuzzy character that is essential to the application of fuzzy processing. Fuzzy set theory considers uncertainty in the form of a membership function, and therefore fuzzy sets can process imperfect data if this imperfection originates from vagueness and ambiguity rather than randomness. Fuzzy contrast enhancement can improve edge detection and, by extension, the quality of related classification features. In this paper, classical (Linguistic hedges and fuzzy enhancement functions), advanced fuzzy sets (Intuitionistic fuzzy set (ΙFS), Pythagorean fuzzy set (PFS), and Fermatean fuzzy sets (FFS)), and a Genetic Algorithm optimizer are proposed to enhance the contrast of mammographic features. The advanced fuzzy sets provide better information on the uncertainty of the membership function. As a result, the intuitionistic method had the best overall performance, but most of the techniques could be used efficiently, depending on the problem that needed to be solved. Linguistic methods could provide a more manageable way of spreading the histogram, revealing more extreme values than the conventional methods. A fusion technique of the enhanced mammography images with Ordered Weighted Average operators (OWA) achieves a good-quality final image. Full article

In this paper, a fuzzy-logic-based energy management system and a multi-port quasi-z-source converter that utilizes hybrid renewable energy sources are proposed. The system ensures that each energy source module can be used individually by employing fuzzy logic to define the power modes. This approach also helps to prevent switching losses resulting from the extra switching of the source modules. In addition, the proposed energy management does not have a mathematical model, so its applicability is simple, and it is suitable for different multiple-input topologies. The Mamdani fuzzy inference system can be designed to capture the nonlinear behavior of the system owing to linguistic rules. Moreover, the switching losses of the multiport modules were significantly reduced by adopting the quasi-z-source network to the end of the multiport converter. Furthermore, different errors, such as the root mean square error (RMSE), average squared error (ASE), average absolute error (AAE), average time-weighted absolute error (ATWAE), tracking error (TE), and unscaled mean bounded relative absolute error (UMBRAE), were applied to evaluate the fuzzy logic performance from different perspectives. The simulation results were obtained using MATLAB Simulink, and the experimental results were obtained by connecting the circuit to MATLAB Simulink using an Arduino Due. Full article

Complex fuzzy sets (CFSs) are a suitable tool to manage spatial directional information which includes distance and direction. However, spatial directional information is given by linguistic values. It is very awkward for the CFS to describe this type of spatial directional information. To overcome this limitation, we first propose a novel concept called a linguistic complex fuzzy set (LCFS) to serve as an extension of the CFS. Then we put forward some basic operational laws for LCFSs. After that, we define three operators for LCFSs: the linguistic complex fuzzy weighted averaging (LCFWA) operator, the linguistic amplitude max (Amax) operator and the linguistic amplitude min (Amin) operator. In actual application, we use the LCFWA operator to deal with group decision making when the importance weights of experts are known. For the situation in which the weights of experts are unknown, we develop an Amax-Amin method for group decision making. Full article

In fuzzy artificial intelligent decision support systems, three-way intelligent-decision making (TWIDM) has played a very important role in ranking objects under the double hierarchy linguistic variable (DHLV). The 8 × 8 S-boxes are very important for image encryption in secure communication. Therefore, the aim of the present study is to develop a new approach to artificial intelligent three-way decision making via DHLV and apply it to S-box image encryption. Artificial intelligent based three-way decision-making problems with double hierarchy hesitant linguistic terms are developed. The first and second hierarchy hesitant linguistic term sets make up the double hierarchy hesitant linguistic term set, which allows for more flexible expressions of doubt and fuzziness. First, we define the Einstein operational laws, score function, and Einstein aggregation operators; i.e., double hierarchy hesitant linguistic Einstein weighted averaging and weighted geometric operators. First, the unknown weight vector for decision experts is determined by using aggregation operators and entropy measures for DHLV. Then, we find the weight vector for our criteria by using the distance measure. In TWIDM, conditional probability is determined by using the extended TOPSIS method for evaluating the S-boxes for image encryption. The expected losses are then computed by aggregating the loss functions with the help of Einstein-weighted averaging aggregation operators. Finally, we apply the minimum-loss decision rules for the selection of S-box to image encryption. The proposed decision technique has been compared with existing three-way decisions and the result of proposed three-way decision making for analyzing and ranking the S-box is very good and reliable for decision making. Full article

The aim of remote sensing image captioning (RSIC) is to describe a given remote sensing image (RSI) using coherent sentences. Most existing attention-based methods model the coherence through an LSTM-based decoder, which dynamically infers a word vector from preceding sentences. However, these methods are indirectly guided through the confusion of attentive regions, as (1) the weighted average in the attention mechanism distracts the word vector from capturing pertinent visual regions and (2) there are few constraints or rewards for learning long-range transitions. In this paper, we propose a multi-source interactive stair attention mechanism that separately models the semantics of preceding sentences and visual regions of interest. Specifically, the multi-source interaction takes previous semantic vectors as queries and applies an attention mechanism on regional features to acquire the next word vector, which reduces immediate hesitation by considering linguistics. The stair attention divides the attentive weights into three levels—that is, the core region, the surrounding region, and other regions—and all regions in the search scope are focused on differently. Then, a CIDEr-based reward reinforcement learning is devised, in order to enhance the quality of the generated sentences. Comprehensive experiments on widely used benchmarks (i.e., the Sydney-Captions, UCM-Captions, and RSICD data sets) demonstrate the superiority of the proposed model over state-of-the-art models, in terms of its coherence, while maintaining high accuracy. Full article

In this research, we provide tools to overcome the information loss limitation resulting from the requirement to estimate the results in the discrete initial expression domain. Through the use of 2-tuples, which are made up of a linguistic term and a numerical value calculated between $[0.5,0.5)$, the linguistic information will be expressed. This model supports continuous representation of the linguistic data within its scope, permitting it to express any information counting received through an aggregation procedure. This study provides a novel approach to develop a linguistic multi-attribute group decision-making (MAGDM) approach with complex fractional orthotriple fuzzy 2-tuple linguistic (CFOF2TL) assessment details. Initially, the concept of a complex fractional orthotriple fuzzy 2-tuple linguistic set (CFO2TLS) is proposed to convey uncertain and fuzzy information. In the meantime, simple aggregation operators, such as CFOF2TL weighted average and geometric operators, are defined. In addition, the CFOF2TL Maclaurin’s symmetric mean (CFOF2TLMSM) operators and their weighted shapes are presented, and their attractive characteristics are also discussed. A new MAGDM approach is built using the developed aggregation operators to address managing economic crises under COVID-19 with the CFOF2TL information. As a result, the effectiveness and robustness of the developed method are accompanied by an empirical example, and a comparative study is carried out by contrasting it with previous approaches. Full article

The linguistic interval-valued Pythagorean fuzzy (LIVPF) sets, which absorb the advantages of linguistic terms set and interval-valued Pythagorean fuzzy sets, can efficiently describe decision makers’ evaluation information in multi-attribute group decision-making (MAGDM) problems. When investigating aggregation operators of linguistic interval-valued Pythagorean fuzzy (LIVPF) information, we have to consider two important issues, viz. the operational rules of LIVPF numbers and aggregation functions. The classical Archimedean t-norm and t-conorm (ATT) are a famous t-norm and t-conorm, which can produce some special cases. Recently, ATT has been widely applied in different fuzzy decision-making information. Hence, in this paper, for the first issue, we propose some novel operational rules of LIVPF numbers based on ATT. The new operational laws are flexible and can generate some useful operations. For the second issue, we choose a powerful function, i.e., the extended power average (EPA) operator as the aggregation function. The prominent advantages of EPA are that it not only considers the relationship among input arguments, but also dynamically changes the weights of input arguments by employing a parameter. Hence, our proposed novel aggregation operators for LIVPFNs are flexible and is suitable to handle MAGDM problems in actual life. Afterward, we further present a novel MAGDM method under LIVPF conditions. The main finding of our study is a new MAGDM method, which is more powerful and flexible than existing ones. Finally, we apply the method in a sustainable building materials selection to show its effectiveness. Additionally, comparison analysis is provided to demonstrate the advantages and superiorities of the proposed method. Full article