Data Science and Management最新文献

{"title":"Cross-domain aspect term extraction using pre-trained language models with pre-training and fine-tuning strategy","authors":"Chunjun Zhao ,&nbsp;Meiling Wu ,&nbsp;Xuzhuang Sun ,&nbsp;Lu Kang","doi":"10.1016/j.dsm.2025.08.003","DOIUrl":"10.1016/j.dsm.2025.08.003","url":null,"abstract":"<div><div>As an important subtask of fine-grained sentiment analysis, aspect term extraction (ATE) aims to identify aspect terms within user-generated comments. ATE supervised learning approaches are heavily based on the availability of annotated data with token-level labels. However, obtaining these annotations for each domain in sufficient quantity is often a costly process that limits the applicability of supervised methods. The cross-domain aspect term extraction (CDATE) problem involves transferring knowledge from a well-annotated source domain to a less-annotated target domain. We propose an approach combining the pre-trained language model with the pre-training and fine-tuning strategy for solving CDATE (named LM-PF). The first stage adapts bidirectional encoder representations from transformers (BERT) using unlabeled data from both domains to capture domain-specific characteristics, the second stage fine-tunes a downstream sequence labeling model with minimal labeled data to improve target domain performance, the third stage pre-trains on source-domain data and fine-tunes the bidirectional long short-term memory with conditional random fields (Bi-LSTM+CRF) with limited target labels. Specifically, we first adapt a pre-trained BERT model using unlabeled data from both the source and target domains to better capture domain-specific language patterns. Then, we build an ATE model initialized with this adapted BERT and pre-train it using labeled source domain data. Finally, we fine-tune the model on labeled target domain data to enhance its generalization to new domains. We conducted six CDATE tasks on three benchmark datasets, restaurants, laptops, and digital devices. The results show that our model achieves the highest average Micro-F1 score (60.09%) across all tasks and outperforms strong baselines such as generative cross-domain data augmentation by an average margin of 2.7%, confirming the effectiveness of combining domain-adaptive pretraining with task-specific fine-tuning in enhancing cross-domain generalization for ATE tasks.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"9 1","pages":"Article 100163"},"PeriodicalIF":0.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147549814","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Unveiling the footprints of explainable AI in Industry 4.0/5.0: a systematic review and bibliometric exploration","authors":"Oladosu Oyebisi Oladimeji ,&nbsp;Gbadegesin Adetayo Taiwo ,&nbsp;Dauda Adeite Adenusi","doi":"10.1016/j.dsm.2025.07.002","DOIUrl":"10.1016/j.dsm.2025.07.002","url":null,"abstract":"<div><div>Progress in artificial intelligence (AI) is driving transformations that compel an increasing number of companies to embrace the Industry 4.0 and 5.0 paradigms and adopt advanced AI solutions. However, the lack of explainability in evaluating and justifying the decisions of AI models raises significant concerns regarding transparency, safety, and regulatory compliance, particularly in critical industrial applications. Hence, to understand the current explainable AI (XAI) approaches in the industry, this study performed a bibliometric analysis and systematic review concentrating on XAI within industrial research. In addition, the current trends address challenges and outline future directions. This study provides valuable insights that can aid researchers and practitioners in understanding the primary challenges involved in the effective application of XAI in Industry 4.0/5.0 research. Using the Scopus database, 82 articles published between 2019 and 2023 were analyzed, with an emphasis on the relevant outlets, institutions, countries, and keywords. Based on the review findings, there has been a continuous growth in XAI in industrial research. Future research should emphasize human-centric design, interdisciplinary collaboration, and continuous learning to enhance the adoption and effectiveness of XAI in Industry 4.0/5.0. This analysis shows that XAI in industry research is still in its infancy but tends to grow in the future. This review not only maps academic developments but also highlights the implications of XAI for industrial decision-making and managerial practice in smart manufacturing and related domains.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"9 1","pages":"Article 100160"},"PeriodicalIF":0.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147549813","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Artificial intelligence dimensions and design features for the knowledge-based work of the future: a socio-technical perspective","authors":"Jungwoo Lee ,&nbsp;Jaehyun Park ,&nbsp;Noha A. Mostafa","doi":"10.1016/j.dsm.2025.05.002","DOIUrl":"10.1016/j.dsm.2025.05.002","url":null,"abstract":"<div><div>With recent advances in transformative technologies, such as the Internet of Things, cloud computing, and mobile devices, jobs have become more knowledge-based. To implement successful technologies in any society, user perception and acceptance are critical. This study aims to identify a range of artificial intelligence (AI) design features applicable to future-oriented, knowledge-based work from a socio-technical perspective. Industrial reports on AI design artifacts appearing from 2015 to 2025 were reviewed to understand how AI can revolutionize business process management by integrating data analytics, automation, and real-time insights. Subsequently, a typological framework for AI design features was proposed, focusing on automation and intelligence as the two primary dimensions. Using this framework, nine AI design features were derived and discussed. A design feature matrix concerning the changing nature of work was constructed using nine AI design features, and the characteristics of each design feature were described in detail for emerging industry cases. The implications of this framework are discussed in terms of future knowledge, socio-materiality, and expected AI-design affordances. The primary contribution of this study is the proposal of a framework of AI design features for knowledge-based future work from a socio-technical perspective, paving the way for Industry 5.0.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"9 1","pages":"Article 100151"},"PeriodicalIF":0.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147657821","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Introducing article numbering to Data Science and Management","authors":"","doi":"10.1016/S2666-7649(26)00019-6","DOIUrl":"10.1016/S2666-7649(26)00019-6","url":null,"abstract":"","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"9 1","pages":"Article 100189"},"PeriodicalIF":0.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147951173","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Effect of digital ecosystem embeddedness on farmers’ entrepreneurial performance: evidence from China","authors":"Yuhua Cao ,&nbsp;Yunge Fan ,&nbsp;Yumei Wang ,&nbsp;Xinyu Liao","doi":"10.1016/j.dsm.2025.06.004","DOIUrl":"10.1016/j.dsm.2025.06.004","url":null,"abstract":"<div><div>Digital ecosystem embeddedness subverts the promotion mechanism of farmers’ entrepreneurial performance; however, related research is scarce. This study constructed a theoretical mechanism model and applied a structural equation model, regression analysis, and bootstrapping to explore the impact of digital ecosystem embeddedness on farmers’ entrepreneurial performance, based on survey data from 592 farmer start-ups in South China and East China. The research finds that digital ecosystem embeddedness has a significant impact on social networks and farmers’ entrepreneurial performance, with social networks acting as a mediator between digital ecosystem embeddedness and farmers’ entrepreneurial performance. Social networks positively affect entrepreneurial learning and resource management, and entrepreneurial learning and resource management positively affect farmers’ entrepreneurial performance. Both entrepreneurial learning and entrepreneurial resource management exert a mediating effect between social network and farmers’ entrepreneurial performance, respectively. Digitalization capability exerts a moderating effect between digital ecosystem embeddedness and social networks. The results show that the promotion mechanism of digital ecosystem embeddedness on farmers’ entrepreneurial performance constructs a new cognitive and resource space for entrepreneurs by rebuilding social networks to provide a combination of continuous learning support and efficient resource aggregation. This study enriches our understanding of the effect of digital ecosystem embeddedness on farmers’ entrepreneurial performance by depicting an integrated mechanism for entrepreneurial performance promotion in a digital context. It also provides practical guidance for farming start-ups to develop sustainable entrepreneurial advantages in a digital context.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"9 1","pages":"Article 100156"},"PeriodicalIF":0.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147549815","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enabling innovation with big data analytics capabilities: the moderating role of organizational learning culture","authors":"Omar Mohammed Horani ,&nbsp;Ahmad Samed Al-Adwan ,&nbsp;Waleed Mugahed Al-Rahmi ,&nbsp;Ali Alkhalifah","doi":"10.1016/j.dsm.2025.06.006","DOIUrl":"10.1016/j.dsm.2025.06.006","url":null,"abstract":"<div><div>Big data analytics (BDA) capabilities for innovation are an area of growing interest; however, empirical results on this pivotal relationship remain inconclusive. This study investigates how BDA capabilities influence product, process, and business model innovation while examining the moderating role of organizational learning culture (OLC). Drawing on the dynamic capabilities perspective, we conceptualize BDA capabilities in four dimensions: data-driven culture, technological infrastructure, human analytical skills, and data governance. Using survey data from 298 manufacturing firms in Jordan and employing partial least squares structural equation modeling, we find that these competencies significantly enhance all three types of innovation, with their effects further amplified by a strong learning culture. Our findings extend prior research by revealing the specific mechanisms through which data-driven strategies foster innovation and highlighting OLC as a critical social integration mechanism. This study contributes to the understanding of how data-centric approaches reshape industry dynamics and offers actionable insights for firms seeking innovation-led competitiveness.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"9 1","pages":"Article 100158"},"PeriodicalIF":0.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147657820","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Efficient management of medical image datasets for content-based case retrieval via sparse annotation and bidirectional label propagation","authors":"Chengyue Gong ,&nbsp;Ziyu Guan ,&nbsp;Jianguo Ju","doi":"10.1016/j.dsm.2025.07.001","DOIUrl":"10.1016/j.dsm.2025.07.001","url":null,"abstract":"<div><div>Efficient management of medical image datasets is critical for clinical decision-making. However, current methods lack fine-grained annotation management and content-based case retrieval capabilities. Existing systems struggle to handle the annotation burden and to efficiently retrieve similar cases based on image content. To address these challenges, we propose a novel method that uses sparse annotation at the beginning, middle, and end of key disease regions, thereby reducing annotation effort while maintaining segmentation accuracy. We introduce CA-Morpher, an unsupervised image registration model enhanced by a cross-attention mechanism that effectively propagates sparse labels. Additionally, we develop a bidirectional label transfer algorithm (BLTA) that combines prior annotations and registration to propagate labels through bidirectional transfer and pseudo-label weighted fusion. This approach improves the management of annotated datasets by enabling content-based case retrieval, thereby enhancing overall data management efficiency. Experimental results show that our method achieves a Dice score of 76.62%, a Jaccard index of 63.25%, and a Hausdorff distance (HD) of 4.64 on the medical segmentation decathlon (MSD) pancreatic tumor dataset, outperforming current sparse annotation methods and significantly improving the management and retrieval of medical image data.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"9 1","pages":"Article 100159"},"PeriodicalIF":0.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147657819","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multilingual opinion mining of imbalanced customer review datasets using GPT and ensemble hybrid deep learning model","authors":"Nouri Hicham ,&nbsp;Habbat Nassera","doi":"10.1016/j.dsm.2025.06.003","DOIUrl":"10.1016/j.dsm.2025.06.003","url":null,"abstract":"<div><div>This study introduces stacked deep learning for multilingual opinion mining. This framework incorporates RoBERTa-GRU, RoBERTa-LSTM, RoBERTa-BiGRU, and RoBERTa-BiLSTM hybrid models, optimized using the Adam optimizer. The methodology can handle three languages: French, English, and Arabic. This research explores the challenges of imbalanced datasets in opinion mining, employing oversampling approaches like advanced easier data augmentation, synthetic minority over-sampling technique, and generative pre-trained transformer to balance the datasets and improve classification efficiency. We used the assessment metrics of Cohen’s kappa, the receiver operating characteristic-area under curve, accuracy, and mattheus correlation coefficient, along with <em>k</em>-fold validation, to evaluate the sentiment analysis performance across three languages and six datasets. Moreover, we computed performance metrics for all models while scaling the dataset for training and testing. We also determined the memory usage and execution time for each model. Utilizing a stacked deep learning algorithm, the suggested methodology for multilingual opinion mining demonstrated high efficacy in extracting meaningful insights from social media data across several languages. The technique produced significant outcomes, rendering it a potentially valuable instrument for enhancing performance and customer satisfaction by identifying patterns and trends in public sentiment.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"9 1","pages":"Article 100155"},"PeriodicalIF":0.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147657885","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Impact of China's outward foreign direct investment on the Belt and Road Initiative countries' digital economy and energy transition","authors":"Jiahui Jiang ,&nbsp;Pin Li ,&nbsp;Jinsuo Zhang","doi":"10.1016/j.dsm.2025.01.003","DOIUrl":"10.1016/j.dsm.2025.01.003","url":null,"abstract":"<div><div>Digital economy (DE) is on the verge of becoming a pivotal facilitator of the energy transition (ET). However, it remains to be seen to what extent China's outward foreign direct investment (OFDI) can stimulate the development of the DE and promote ET under the Belt and Road Initiative (BRI). Employing a nonparametric additive model, this study examines this issue by using panel data from 84 BRI countries spanning 2007–2022. The connections between China's OFDI and the DE and ET of BRI nations vary across different regions. China's OFDI has considerably fostered the ET of nations along its route to Central Asia, Europe, North Africa, and the Middle East. Conversely, in the Pacific and East Asian regions, China's OFDI has a constructive effect on the advancement of DE in participating nations. However, the effect of China's OFDI on the ET of nations in the Pacific Islands and East Asia and South Asia is statistically insignificant. Moreover, China's OFDI impairs the DE of nations in Central Asia and Europe, the Caribbean and Latin America, as well as North Africa and the Middle East. Notably, the net effect of China's OFDI on ET is positive; however, the mediating effect of facilitating ET through DE is not pronounced.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"9 1","pages":"Article 100138"},"PeriodicalIF":0.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147549798","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A multi-location inventory system with occasional allocation under uncertain defective rates","authors":"Ligang Cui ,&nbsp;Zuguo Peng ,&nbsp;Yingcong Chen ,&nbsp;Wenjing Xie","doi":"10.1016/j.dsm.2025.05.001","DOIUrl":"10.1016/j.dsm.2025.05.001","url":null,"abstract":"<div><div>In this study, a multi-location allocation problem is investigated under a temporary allocation system in light of the uncertain defective rate inherent to items. Based on the theoretical and modeling foundation of previous research, this study contributes in four main aspects. First, a robust optimization model is developed based on the budget uncertainty of the defective rate of items in a multi-location inventory system. Second, given that the determination of uncertainty relies on the subjective judgment of decision-makers with different risk preferences, a pivotal variable method is introduced to approximate the possible thresholds of the uncertain parameter within the expected interval. Third, a new evaluation criterion curve, illustrating the cumulative cost differences between the probabilistic and robust models, is employed through large-scale sampling to evaluate the performance of the two decision models. Finally, numerical experiments and sensitivity analysis are conducted to verify the performance of the decision effects under different defective rates and stockout costs, respectively. The results confirm the superior performance of the robust model, as shown through sensitivity analysis under thresholds of varying defective rates and stockout costs, and highlight the effectiveness of the proposed cumulative cost difference evaluation curve. The robust model offers significant cost-control advantages over the probabilistic model, particularly in high-uncertainty scenarios. This study provides a reference method for multi-location inventory allocation decisions with uncertain defective rates.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"9 1","pages":"Article 100150"},"PeriodicalIF":0.0,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147549812","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}