Search results

Aspect category-based sentiment analysis (ACSA) has been widely used in consumer preference mining and marketing strategy formulation. However, existing studies ignore the variability in features and the intrinsic correlation among diverse aspect categories in ACSA tasks. To address these problems, this paper aims to propose a novel integrated framework.

The integrated framework consists of three modules: text feature extraction and fusion, adaptive feature selection and category-aware decision fusion. First, text features from global and local views are extracted and fused to comprehensively capture the potential information in the different dimensions of the review text. Then, an adaptive feature selection strategy is devised for each aspect category to determine the optimal feature set. Finally, considering the intrinsic associations between aspect categories, a category-aware decision fusion strategy is constructed to enhance the performance of ACSA tasks.

Comparative experimental results demonstrate that the integrated framework can effectively detect aspect categories and their corresponding sentiment polarities from review texts, achieving a macroaveraged F1 score (Fmacro) of 72.38% and a weighted F1 score (F1) of 79.39%, with absolute gains of 2.93% to 27.36% and 4.35% to 20.36%, respectively, compared to the baselines.

This framework can simultaneously detect aspect categories and corresponding sentiment polarities from review texts, thereby assisting e-commerce enterprises in gaining insights into consumer preferences, prioritizing product improvements, and adjusting marketing strategies.

The objective of this study is to accurately predict the cost of green buildings to provide quantifiable criteria for investment decisions from investors.

This study proposes a hybrid prediction model ML-based for cost prediction of GBPs and obtains prediction parameters (PPs) associated with project characteristics through data mining (DM) techniques. The model integrates a principal component analysis (PCA) method to perform parameter dimensionality reduction (PDR) on a large number of raw variables to provide independent characteristic terms. Moreover, the support vector machine (SVM) algorithm is improved to optimize the prediction results and integrated with parameter dimensionality reduction and cost prediction.

The prediction results show that the mean absolute and relative errors of the hybrid prediction model proposed in this study are equal to 39.78 and 0.02, respectively, which are much lower than those of the traditional SVM model and MRA prediction model. Moreover, the hybrid prediction model with parameter dimensionality reduction also achieved better prediction accuracy (R² = 0.319) and superior prediction accuracy for different cost terms.

Theoretically, the hybrid prediction model developed in this study can reliably predict the cost while accurately capturing the characteristics of GBPs, which is a bold attempt at a comprehensive approach. Practically, this study provides developers with a new ML-based prediction model that is capable of capturing the costs of projects with ambiguous definitions and complex characteristics.

The purpose of this study is to map the intellectual structure of sustainable finance and accounting (SFA) literature by identifying the influential aspects, main research streams and future research directions in SFA.

The results are obtained using bibliometric citation analysis and content analysis to conduct a bibliometric review of the intersection of sustainable finance and sustainable accounting using a sample of 795 articles published between 1991 and November 2023.

The most influential factors in the SFA literature are identified, highlighting three primary areas of research: corporate social responsibility and environmental disclosure; financial and economic performance; and regulations and standards.

SFA has experienced rapid development in recent years. The results identify the current research domain, guide potential future research directions, serve as a reference for SFA and provide inspiration to policymakers.

SFA typically encompasses sustainable corporate business practices and investments. This study contributes to broader social impacts by promoting improved corporate practices and sustainability.

This study expands on previous research on SFA. The authors identify significant aspects of the SFA literature, such as the most studied nations, leading journals, authors and trending publications. In addition, the authors provide an overview of the three major streams of the SFA literature and propose various potential future research directions, inspiring both academic research and policymaking.

This study quantitatively investigates the impacts of digital and learning orientations on supply chain resilience (SCR) and firm performance (FP), aiming to fill the gaps in understanding their specific impacts in the context of Industry 4.0 developments and supply chain disruptions.

This study utilized survey techniques and structural equation modelling (SEM) to gather and analyse data through a questionnaire based on a seven-point Likert scale. Hypotheses were formulated based on an extensive literature review and tested using Amos software.

The study confirms SCR’s significant impact on FP, aligning with existing research on resilience’s role in organizational competitiveness. This study uncovers the nuanced impacts of digital and learning orientations on SCR and FP. Internal digital orientation (DOI) positively impacts SCR, while external digital orientation (DOE) does not. Specific dimensions of learning orientation – shared vision (LOS), open-mindedness (LOO) and intraorganizational knowledge sharing (LOI) – enhance SCR, while commitment to learning (LOC) does not. SCR mediates the relationship between DOI and FP but not between DOE and FP.

This research focuses on digital and learning orientations, recommending that future studies investigate other strategic orientations and examine the specific contributions of various digital technologies to SCR across diverse contexts.

The empirical findings emphasize the significance of developing internal digital capabilities and specific learning orientations to enhance SCR and FP, aligning these initiatives with resilience strategies.

This study advances knowledge by distinguishing the impacts of internal and external digital orientations and specific learning dimensions on SCR and FP, offering nuanced insights and empirical validation.

With the development of Web information systems, steel e-commerce platforms have accumulated a large number of quality objection texts. These texts reflect consumer dissatisfaction with the dimensions, appearance and performance of steel products, providing valuable insights for product improvement and consumer decision-making. Currently, mainstream solutions rely on pre-trained models, but their performance on domain-specific data sets and few-shot data sets is not satisfactory. This paper aims to address these challenges by proposing more effective methods for improving model performance on these specialized data sets.

This paper presents a method on the basis of in-domain pre-training, bidirectional encoder representation from Transformers (BERT) and prompt learning. Specifically, a domain-specific unsupervised data set is introduced into the BERT model for in-domain pre-training, enabling the model to better understand specific language patterns in the steel e-commerce industry, enhancing the model’s generalization capability; the incorporation of prompt learning into the BERT model enhances attention to sentence context, improving classification performance on few-shot data sets.

Through experimental evaluation, this method demonstrates superior performance on the quality objection data set, achieving a Macro-F1 score of 93.32%. Additionally, ablation experiments further validate the significant advantages of in-domain pre-training and prompt learning in enhancing model performance.

This study clearly demonstrates the value of the new method in improving the classification of quality objection texts for steel products. The findings of this study offer practical insights for product improvement in the steel industry and provide new directions for future research on few-shot learning and domain-specific models, with potential applications in other fields.

This study provides a comprehensive overview of the research landscape of Kansei engineering (KE) within the domain of emotional clothing design. It explores the pivotal technologies, challenges and potential future directions of KE, offering application methodologies and theoretical underpinnings to support emotional clothing design.

This study briefly introduces KE, outlining its overarching research methodologies and processes. This framework lays the groundwork for advancing research in clothing Kansei. Subsequently, by reviewing literature from both domestic and international sources, this research initially explores the application of KE in the design and evaluation of clothing products as well as the development of intelligent clothing design systems from the vantage point of designers. Second, it investigates the role of KE in the customization of online clothing recommendation systems and the optimization of retail environments, as perceived by consumers. Finally, with the research methodologies of KE as a focal point, this paper discusses the principal challenges and opportunities currently confronting the field of clothing Kansei research.

At present, studies in the domain of clothing KE have achieved partial progress, but there are still some challenges to be solved in the concept, technical methods and area of application. In the future, multimodal and multisensory user Kansei acquisition, multidimensional product deconstruction, artificial intelligence (AI) enabling KE research and clothing sales environment Kansei design will become new development trends.

This study provides significant directions and concepts in the technology, methods and application types of KE, which is helpful to better apply KE to emotional clothing design.

Plant-based meat analogues (PBMAs) hold significant promise as a sustainable solution to meet future protein demands, replicating the taste and nutritional value of meat. However, the present reliance on extrusion technology in PBMA production limits the exploration of more accessible and affordable methods. The current investigation aims to meet the market demand for a scalable and cost-effective processing approach by exploring saturated steam-assisted technology that could broaden the production volume of PBMAs, thereby supplementing protein security and planet sustainability.

A one-factor-at-a-time (OFAT) approach is employed to evaluate the effect of ingredients and process conditions on the governing quality attributes (texture, colour and sensory).

Among the ingredients, monosodium glutamate (MSG) and nutritional yeast (NY) significantly enhanced the hardness and chewiness of saturated steam-assisted plant-based meat analogues (ssPBMAs) followed by potato protein isolate (PPI), defatted soy flour (DSF) and salt. The addition of PPI and DSF led to a decrease in lightness (L* value) and an increase in the browning index (BI). Sensory evaluations revealed that higher concentrations of DSF imparted a noticeable beany flavour (>20%), whereas PPI (30%) improved the overall sensory appeal. Increased levels of NY (10%) and MSG (5%) enhanced the umami flavour, enhancing consumer preference. Higher thermal exposure time (TTi) (45 min) and temperature (TTe) (120 °C) during processing resulted in softer products with reduced L* values. These findings establish a foundation for selecting and optimizing the ingredients and processing parameters in ssPBMA production.

The novelty of the current study includes process behaviour of selected ingredients such as PPI, NY, MSG, DSF, salt and adopted process conditions, namely, dough processing time (DPT), protein network development time (PNDT), TTi and TTe on the quality of ssPBMAs.

This study aims to efficiently estimate the extremely small failure probability with high-dimensional inputs and multiple failure domains.

This paper proposed an adaptive stratified mixture importance sampling method. The proposed method first constructs an explicit and regular mixture importance sampling probability density function (M-IS-PDF) by taking the clustering centroids as the density centers. Then by the constructed M-IS-PDF, the proposed method explores the rare multiple failure domains by adaptively stratifying, thereby addressing the issue of estimating extremely small failure probability robustly and efficiently.

Compared with the existing cross-entropy based IS method, the constructed M-IS-PDF not only covers the domains significantly contributing to the failure probability through clustering centroids to reduce the variance of failure probability estimation, but also has no undetermined parameter set to optimize, enhancing the adaptability in high-dimensional problems. Compared with the subset simulation method, the adaptive stratified M-IS-PDF constructed is explicit, regular and easy sampling. It not only has high sampling efficiency but also avoids estimating conditional failure probabilities layer by layer, improving the algorithmic robustness for estimating extremely small failure probability.

Both numerical and engineering examples indicate that, under the similar failure probability estimation accuracy, the proposed method requires significantly smaller sample size and lower computational cost than subset simulation and cross-entropy based IS methods, demonstrating higher efficiency and robustness in addressing intractable reliability analysis problems with high-dimensional inputs, multiple failure domains and rare failure.

Supply chains, as prototypical uncertain systems, are crucial for national security and socioeconomic development. Grey system theory (GST) is an effective tool for addressing uncertainties and has played a pivotal role in related research within the supply chain domain. This study aims to systematically summarize the research achievements and knowledge structures pertaining to GST in supply chain studies. Current and potential research hotspots are also analyzed.

CiteSpace is used to conduct a bibliometric analysis of 1,617 articles sourced from the Web of Science (WOS). The analysis aims to summarize the current state of research and the knowledge structure in the field. A strategic diagram incorporating two data indicators, namely, novelty and concern, is constructed based on keyword clustering to identify and analyze current and potential research hotspots.

Studies utilizing GST to guide supply chain practices have attracted the interest of scholars from 205 research institutions across 85 countries and regions globally, which earned recognition from 183 high-level academic journals. In this field, the School of Economics and Management at Nanjing University of Aeronautics and Astronautics stands out as a core research institution, with Professor Deng Julong, who is the founder of GST, being the most frequently cited scholar. Current research hotspots are complex equipment supply chains, drivers and challenges in supply chain management, supply chain risk management, closed-loop supply chain and supply chain operation in the big data era. In addition, emerging research hotspots include digital and intelligent logistics technology, sustainable supplier management, determinants and flexibility of supply chain contracts, supply chain strategy, purchase management, grey prediction of demand and consumption, grey forecasting and economy efficiency, China-specific issues and grey model construction.

The bibliometric analysis reveals the current state and knowledge structure of research in this field. Previous studies on research hotspots have primarily relied on subjective judgments, which lacked empirical data support. This study constructs a strategic diagram incorporating two data indicators, namely, novelty and concern, to provide a more objective and reliable analysis of research hotspots.