Electronic Commerce Research and Applications最新文献

{"title":"Uncovering financial statement fraud detection with multi-modal positive-unlabeled contrastive learning","authors":"Xusheng Sun ,&nbsp;Jingli Huang ,&nbsp;Gang Wang ,&nbsp;Xuan Zhang ,&nbsp;Manlu Liu ,&nbsp;Shanlin Yang","doi":"10.1016/j.elerap.2026.101604","DOIUrl":"10.1016/j.elerap.2026.101604","url":null,"abstract":"<div><div>Current research in AI-based financial statement fraud detection (FSFD) predominantly draws on financial and textual modalities and employs supervised machine learning. However, few attend to the inconsistency between these modalities, which can provide a critical indicator in FSFD. Moreover, the assumption underlying supervised learning is often violated, as non-fraudulent at present may later be revealed to exhibit fraudulent behavior. Accordingly, we introduce multi-modal inconsistency learning combined with Positive-Unlabeled (PU) learning, which trains a classifier using only positive and unlabeled data. Nevertheless, two challenges persist: the semantic gap between modalities in inconsistency learning and selection bias when identifying reliable negative instances in PU learning. To overcome these challenges, we propose PUMLCL: Positive-Unlabeled learning with Multi-Level Contrastive Learning. First, to bridge the semantic gap between modalities, we employ instance-level contrastive learning (CL) and introduce an inconsistency-enhancing module for the multi-modal inconsistency learning. Simultaneously, we introduce industry-level contrastive learning to learn the representations that capture distinctive industrial characteristics. Next, utilizing the learned industry-aware representations, we mitigate selection bias by identifying reliable negative instances with respect to the positive ones across industries. Finally, we design a classifier augmented with group-level contrastive loss to enhance the separability between fraudulent and truthful companies. Extensive experiments on real-world datasets validate the effectiveness of PUMLCL.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"77 ","pages":"Article 101604"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147956779","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Corrigendum to “A dynamic data driven-based semi-distributed reputation mechanism in unknown networks” [Electron. Commer. Res. Appl. 14 (2015) 532–541]","authors":"Szu-Yin Lin ,&nbsp;Ping-Hsien Chou","doi":"10.1016/j.elerap.2026.101595","DOIUrl":"10.1016/j.elerap.2026.101595","url":null,"abstract":"","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"77 ","pages":"Article 101595"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148001228","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Do digital nature cues work? Investigating the shaping effect of virtual Influencers’ biophilic design backgrounds on green product preferences","authors":"Xueying Wang ,&nbsp;Yuexian Zhang","doi":"10.1016/j.elerap.2026.101602","DOIUrl":"10.1016/j.elerap.2026.101602","url":null,"abstract":"<div><div>Visual imagery is a critical factor when virtual influencers endorse green products. Given the innate human affinity for nature, incorporating biophilic elements into influencer backgrounds can strengthen consumer associations between green products and environmental benefits, thereby stimulating purchase intentions. However, this relationship remains empirically unexplored. Grounded in the biophilia hypothesis, this study investigates how biophilic design backgrounds in virtual influencer endorsements impact consumer purchase intentions. Across three studies, we find that biophilic backgrounds significantly enhance purchase intentions toward green products, with perceived naturalness and product value serving as mediating mechanisms. Furthermore, the positive effect of perceived naturalness on product value is stronger for hedonic products than for utilitarian products. Additionally, high image-text congruency amplifies the positive impact of biophilic backgrounds on perceived naturalness. This research advances visual marketing literature by offering practical insights for integrating biophilic design into virtual influencer campaigns.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"77 ","pages":"Article 101602"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147850367","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Information technology affordances in online social commerce: conceptual development and empirical examination","authors":"Xueyan Dong,&nbsp;Yuxin Tian,&nbsp;Lin Zhang,&nbsp;Lifang Wang,&nbsp;Tianhui Gao","doi":"10.1016/j.elerap.2026.101601","DOIUrl":"10.1016/j.elerap.2026.101601","url":null,"abstract":"<div><div>Information technology (IT) affordances have garnered widespread interest in online retailing and consumer services literature for theoretically interpreting social media-associated behaviors and for conceptualizing human–computer interaction in IT-integrated environments. However, empirical studies on this phenomenon, especially in the context of online social commerce (OSC), remain limited. This stems partly from the absence of a clear conceptualization of IT affordances and a validated measurement scale within the OSC context. Our study addresses these gaps by conceptualizing IT affordances in OSC and identifying six dimensions of this phenomenon: <em>visibility, metavoicing, social connecting, triggered attending, guidance shopping, and trading</em>. We develop a conceptual model and investigate the relationships between IT affordances and associated user behavior, specifically OSC purchase intention and OSC sharing intention. We further develop and validate a rigorous 26-item scale for IT affordances in OSC through qualitative and quantitative methods. Using data from three rounds of surveys and analyzing them with structural equation modeling, our findings confirm that IT affordances in OSC are strong predictors of OSC purchase and sharing intentions. Our study contributes to online retailing and consumer services research by developing a validated measurement scale for IT affordances in OSC, thereby enabling more precise, context-specific investigations of the role of IT affordances in shaping online retailing and consumer behavior.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"77 ","pages":"Article 101601"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147803236","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Service and pricing strategy based on AI-as-a-service: The value of fine-tuning service","authors":"Shiqiang Yu ,&nbsp;Zhaofang Mao ,&nbsp;Chunxiang Guo","doi":"10.1016/j.elerap.2026.101603","DOIUrl":"10.1016/j.elerap.2026.101603","url":null,"abstract":"<div><div>In e-commerce services, a growing number of service providers are partnering with artificial intelligence (AI) companies to implement AI systems for enhanced service quality. In practice, AI companies offer multiple AI systems and further provide fine-tuning services to address the issue of poor performance in standard systems. We have developed a stylized model to study the optimal strategy for a service provider in selecting AI service combinations. In this model, the service provider pays usage fees and fine-tuning fees to the AI company to obtain AI systems. The results show that in the two-stage scheme, improvements in training effectiveness, reductions in fine-tuning costs, and reductions in data costs do not always benefit the service provider. This is because such changes may incentivize the AI company to increase usage fees, resulting in low profit margins for the service provider. When these parameters fall within moderate ranges, fine-tuning services can significantly enhance profitability and provide benefits to customers. By comparing different AI versions, we find that when the advanced version holds an advantage, an increase in base demand may initially highlight this advantage but subsequently diminish it. Moreover, the four-stage scheme may suppress the AI company’s incentive to increase the usage fee, thereby mitigating the service provider’s loss when the fine-tuning cost coefficient is low. However, this scheme is not always superior to the two-stage scheme. Finally, in the extension, we explore the integration of third-party AI functionalities.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"77 ","pages":"Article 101603"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147859023","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Navigating innovation: the strategic dynamics of AI-driven live streaming","authors":"Haoyu Zhou ,&nbsp;Chaocheng He ,&nbsp;Jiang Wu ,&nbsp;Liuxin Zou","doi":"10.1016/j.elerap.2026.101581","DOIUrl":"10.1016/j.elerap.2026.101581","url":null,"abstract":"<div><div>The rise of AI streamers in the live-streaming industry has significantly reshaped stakeholder strategies and transformed the ecosystem. While these AI streamers offer operational advantages, their adoption also presents notable challenges. Existing research has primarily focused on the business-to-consumer (B2C) perspective, neglecting the broader dynamics of how AI streamers influence multi-stakeholder interactions within the live streaming ecosystem. This study addresses this gap by developing a tripartite evolutionary game model to explore the strategic interactions among brands, Multi-Channel Networks (MCNs), and platforms, thereby providing a comprehensive understanding of AI streamers’ impact on stakeholders’ strategic decisions. The findings indicate that brand adoption of AI streamers is influenced by factors beyond the balance of additional revenue and adoption costs. Even when additional revenue surpasses adoption costs, brands may hesitate due to concerns over long-term sustainability and technological risks. Sensitivity analysis reveals a non-linear relationship between costs and benefits, highlighting that as adoption costs rise, brands may revert to traditional live streaming despite high revenue. Regulatory strategies play a critical role in shaping brand adoption decisions. In particular, moderate supervision fosters both innovation and stability, whereas overly strict or lenient regulations can hinder AI adoption. Thus, effective calibration of regulations supports AI adoption without compromising market stability. This research contributes to the theoretical understanding of multi-stakeholder ecosystems and offers practical insights for integrating AI innovation with effective governance in live-streaming commerce.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"76 ","pages":"Article 101581"},"PeriodicalIF":6.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146173992","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"What makes a review helpful? A multimodal prediction model in e-commerce","authors":"Heena Lim ,&nbsp;Seonu Park ,&nbsp;Qinglong Li ,&nbsp;Xinzhe Li ,&nbsp;Jaekyeong Kim","doi":"10.1016/j.elerap.2026.101586","DOIUrl":"10.1016/j.elerap.2026.101586","url":null,"abstract":"<div><div>The rapid expansion of e-commerce has made online reviews critical to consumer decision-making. However, the overwhelming volume and inconsistent quality of these reviews significantly complicate the identification of valuable information. Consequently, machine learning and deep learning-based review helpfulness prediction (RHP) has emerged as a robust approach for addressing these challenges. While unimodal RHP methods primarily analyze textual content, multimodal approaches enhance predictive accuracy by integrating textual and visual content. Despite these advantages, existing models often overlook shallow features, such as review readability and image quality, which are critical to consumer evaluation. To provide a more comprehensive framework for RHP, an elaboration likelihood model–grounded multimodal cue-based helpfulness-prediction model (MCHPM) is proposed. MCHPM extracts central cues (semantic content from text and images) using bidirectional encoder representations from transformers (BERT) and VGG-16, whereas Python-based feature extraction captures peripheral cues (e.g., text readability and image quality). Finally, a co-attention mechanism identifies cue interdependencies, and a Gated Multimodal Unit effectively weighs each modality during prediction. Empirical evaluation on a real-world Amazon review dataset demonstrates that the proposed MCHPM outperforms benchmark models, thus validating the criticality of incorporating shallow features into multimodal RHP studies. Overall, these findings offer valuable insights into enhancing RHP and ensuring more informed consumer decision-making in e-commerce.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"76 ","pages":"Article 101586"},"PeriodicalIF":6.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147398947","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Key decision criteria for integrating non-fungible tokens (NFTs) into e-commerce platforms","authors":"Pei-Hsuan Tsai ,&nbsp;Silvana Trimi ,&nbsp;Jia-Wei Tang","doi":"10.1016/j.elerap.2026.101575","DOIUrl":"10.1016/j.elerap.2026.101575","url":null,"abstract":"<div><div>This paper develops a comprehensive multi-criteria decision-making (MCDM) framework to evaluate the integration of non-fungible tokens (NFTs) into e-commerce platforms. Grounded in Saaty’s Benefits, Opportunities, Costs, and Risks (BOCR) model, the framework systematically captures both favorable and unfavorable decision dimensions. The Decision-Making Trial and Evaluation Laboratory (DEMATEL) method models causal relationships among the four BOCR dimensions and sixteen evaluation criteria, while the DEMATEL-based Analytic Network Process (DANP) derives their relative weights. Empirical data were obtained through an expert-validated questionnaire completed by 365 e-commerce professionals in Taiwan. DEMATEL results reveal that Cost, Risk, and Opportunity function as causal drivers, whereas Benefit operates as a reactive outcome. However, DANP analysis shows that Benefit holds the highest global weight among the four dimensions, underscoring its centrality in platform-level decision priorities. At the criteria level, Electronic Word of Mouth, System Quality, Data Protection, and Profit-Sharing Mechanisms emerge as most influential. By formalizing the evaluation of interdependent criteria across strategic dimensions, this dual-method framework advances NFT adoption research and offers a replicable, BOCR-based model anchored in platform decision-making. Theoretically, this study contributes by extending BOCR-based evaluation into the emerging context of NFT-enabled commerce and by demonstrating how DEMATEL and DANP can jointly capture causal influence and interdependence among platform-level decision criteria. Practically, the findings provide e-commerce operators with an evidence-based tool for prioritizing system reliability, data protection, incentive mechanisms, and electronic word-of-mouth strategies when designing NFT-augmented services. The proposed framework is also adaptable to broader Web3 innovations, offering firms a structured approach to assessing and comparing token-based business models under conditions of uncertainty.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"76 ","pages":"Article 101575"},"PeriodicalIF":6.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145993602","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"One Size Does Not Fit All: Evaluating the Impact of Bank-FinTech Collaborations on Service Digitalization","authors":"Ting Cao ,&nbsp;Murat Kristal ,&nbsp;Xiaowen Huang","doi":"10.1016/j.elerap.2026.101582","DOIUrl":"10.1016/j.elerap.2026.101582","url":null,"abstract":"<div><div>Digitalization is reshaping retail banking, yet its benefits are not uniform. Banks face the challenges of digitalizing existing services while creating new ones, often in collaboration with FinTech partners. Using survey data from Canadian retail banks and credit unions, this study empirically examines how digitalization efforts, operational agility, and FinTech collaboration jointly influence service performance in the digital age. Results show that digitalizing existing services enhance both standardized and customized services, whereas digitalizing new services benefit only the standardized ones. Moreover, results show that operational agility alone does not deliver value unless combined with FinTech collaboration. This synergy between internal agility and external collaboration is most effective in incremental digitalization of existing services, but not in those more radical development of novel ones. Theoretically, this study extends dynamic capabilities and combinative capabilities perspectives to the context of service digitalization by identifying boundary conditions of digitalization and highlighting the interplay among internal and external capabilities. Practically, this study cautions against the one-size-fits-all digitalization strategies, urging banks to tailor approaches to service types and collaboration choices in their digitalization journey.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"76 ","pages":"Article 101582"},"PeriodicalIF":6.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146174090","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ROQuA: A RAG-based opinion question-answering framework for e-commerce reviews","authors":"Shengkai Zhou ,&nbsp;Jiangtao Qiu ,&nbsp;Ling Lin ,&nbsp;Siyu Wang ,&nbsp;Yun Xu","doi":"10.1016/j.elerap.2026.101576","DOIUrl":"10.1016/j.elerap.2026.101576","url":null,"abstract":"<div><div>With the rapid development of large language models, retrieval-augmented generation (RAG) has become one of the most representative techniques to enhance the ability of LLMs on generating text contents based on retrieved information. However, applying RAG to opinion question–answering (QA) poses two key challenges: (1) retrieving reviews that are relevant to the question and (2) generating opinion summaries without hallucination. In this study, we propose ROQuA, an RAG-based framework designed to address these challenges in opinion QA. The framework incorporates several techniques, including review enrichment, question rewriting, and a three-stage prompting strategy, to improve the performance of ROQuA. We conduct experiments on three datasets—JD, Douyin, and Yelp—containing 182 k, 250 k, and 6 k reviews, respectively. Results demonstrate that ROQuA outperforms state-of-the-art models in opinion QA tasks. In particular, ROQuA achieves a 0.6-point improvement over the second-best model on LSBE, a metric introduced in this study for evaluating opinion QA. In addition, we provide an in-depth analysis of hallucination in RAG-based opinion QA and show that careful review selection and prompt engineering substantially reduce hallucinated content.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"76 ","pages":"Article 101576"},"PeriodicalIF":6.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146025819","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}