Electronic Commerce Research and Applications最新文献

Transparent prediction of financial analyst recommendation quality using generalized additive model 基于广义加性模型的金融分析师推荐质量透明预测

IF 5.9 3区管理学

Electronic Commerce Research and Applications Pub Date : 2025-07-09 DOI: 10.1016/j.elerap.2025.101524

Shuai Jiang , Xiaoxin Pan , Yanhong Guo , Chuanren Liu , Hui Xiong

{"title":"Transparent prediction of financial analyst recommendation quality using generalized additive model","authors":"Shuai Jiang , Xiaoxin Pan , Yanhong Guo , Chuanren Liu , Hui Xiong","doi":"10.1016/j.elerap.2025.101524","DOIUrl":"10.1016/j.elerap.2025.101524","url":null,"abstract":"<div><div>Financial analysts play a key role in financial decision-making, but the reliability of their recommendations can fluctuate dramatically depending on changes in analyst competence and contextual dynamics, posing a significant challenge to investors seeking guidance. This study unveils a novel explainable deep learning architecture, termed Quality Attribution Network (QuANet), which innovates by integrating a Generalized Additive Model framework, amplifying prediction accuracy and facilitating an in-depth understanding of how distinct variables contribute to the quality of analyst recommendations. Further, QuANet incorporates an attention mechanism to discern salient features, thereby ensuring that critical analyst, rating, and stock information receives appropriate weight. Empirical validation on extensive datasets corroborates QuANet’s superiority over existing benchmarks across diverse quality prediction metrics. Enhancing predictive capability translates into tangible gains for investment strategies, underscoring the model’s practical applicability. Additionally, QuANet’s attribution capabilities enable nuanced differentiation between analysts, pinpointing those endowed with genuine expertise within the financial advisory landscape. In sum, this research advances the analytical toolkit for assessing analyst recommendations by introducing a model that harmonizes predictive prowess with interpretative clarity. Investors stand to benefit from the transparent insights generated, facilitating the extraction of valuable knowledge from analyst recommendations to inform judicious investment decisions.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"73 ","pages":"Article 101524"},"PeriodicalIF":5.9,"publicationDate":"2025-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144595731","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}

引用次数: 0

Utilising podcast digital content marketing to influence consumer purchasing behaviour on e-commerce platform: A study on social presence and media richness theories 利用播客数字内容营销影响电子商务平台消费者购买行为：社交存在与媒体丰富度理论研究

IF 5.9 3区管理学

Electronic Commerce Research and Applications Pub Date : 2025-07-04 DOI: 10.1016/j.elerap.2025.101529

Pei-Hsuan Tsai , Ming-Chia Hsieh , Jia-Wei Tang

{"title":"Utilising podcast digital content marketing to influence consumer purchasing behaviour on e-commerce platform: A study on social presence and media richness theories","authors":"Pei-Hsuan Tsai , Ming-Chia Hsieh , Jia-Wei Tang","doi":"10.1016/j.elerap.2025.101529","DOIUrl":"10.1016/j.elerap.2025.101529","url":null,"abstract":"<div><div>Many e-commerce platforms have begun adapting their content management strategies to accommodate the growing popularity of podcasts and attract consumers’ attention. Despite its role in facilitating e-commerce brands to develop their identity and establish field expertise and attracting new consumers from diverse backgrounds, podcasting remains an underused area in the saturated landscape of digital marketing. Hence, this study aims to integrate social presence theory (SPT) and media richness theory (MRT) via grey multiple attribute decision-making (G-MADM) methods to examine the impact of incorporating podcast digital content marketing (DCM) into e-commerce platforms on consumer purchasing behaviour. The current work selected e-commerce platforms in Taiwan, with 656 and 657 valid responses gathered for Study 1 and Study 2, respectively. In examining podcast DCM strategies, causal relationships were determined using grey decision-making and trial evaluation laboratory (G-DEMATEL), while the influence weights of evaluation factors were established using grey DEMATEL-based on analytic network process (G-DANP). The study concludes with the research findings and recommendations. Based on Study 1 (SPT perspective), sense of identity (I) and perceived presence (P) and emotional presence (E) were the key factors requiring improvement. The variety of language (L), immediate feedback (S), and personalisation (H) were the key factors requiring improvement in Study 2 (MRT perspective). Apart from contributing to refining and enriching DCM and sales strategy planning for e-commerce platforms, these findings can facilitate e-commerce platforms to incorporate podcast experiences, improve their DCM strategies, and increase consumer repurchase intent.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"73 ","pages":"Article 101529"},"PeriodicalIF":5.9,"publicationDate":"2025-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144570534","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}

引用次数: 0

How signal intensity of altruistic and strategic motivation affects crowdfunding performance? Matching among funders and platform types 利他动机和战略动机的信号强度如何影响众筹绩效？资助者和平台类型的匹配

IF 5.9 3区管理学

Electronic Commerce Research and Applications Pub Date : 2025-07-02 DOI: 10.1016/j.elerap.2025.101528

Hongke Zhao , Yaxian Wang , Hao Wei

{"title":"How signal intensity of altruistic and strategic motivation affects crowdfunding performance? Matching among funders and platform types","authors":"Hongke Zhao , Yaxian Wang , Hao Wei","doi":"10.1016/j.elerap.2025.101528","DOIUrl":"10.1016/j.elerap.2025.101528","url":null,"abstract":"<div><div>Crowdfunding has gained significant scholarly attention, yet existing research primarily focuses on single-platform studies, limiting the generalizability of findings. We argue that investment motivations vary across platform types, influencing the effectiveness of altruistic and quality signals on crowdfunding performance. Using 114,095 projects from Indiegogo (reward-based) and 1,199,908 loan projects from Kiva (lending-based), we first conduct separate analyses within each platform to examine the impact of these signals. We then compare the marginal effects across platforms to assess how platform structure influences backer decision-making. Our results show that quality signals consistently enhance crowdfunding success but have a stronger influence in reward-based platforms, while the effect of altruistic signals varies, enhancing performance in lending-based platforms but diminishing it in reward-based platforms. Moreover, we identify a reciprocal inhibitory interaction between quality and altruistic signals, suggesting that emphasizing one type of signal may weaken the effectiveness of the other by diverting backers’ attention and influencing how they evaluate the project. These findings underscore the importance of platform differentiation in crowdfunding research and highlight the need to move beyond single-platform studies. Our study offers practical insights for crowdfunding initiators on how to tailor their campaigns based on platform-specific investor behavior.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"73 ","pages":"Article 101528"},"PeriodicalIF":5.9,"publicationDate":"2025-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144557391","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}

引用次数: 0

From knowledge tracing to preference tracing: Capturing dynamic user preferences for personalized recommendation 从知识跟踪到偏好跟踪：捕捉动态用户偏好以进行个性化推荐

IF 5.9 3区管理学

Electronic Commerce Research and Applications Pub Date : 2025-06-30 DOI: 10.1016/j.elerap.2025.101527

Jungmin Hwang , Hakyeon Lee

{"title":"From knowledge tracing to preference tracing: Capturing dynamic user preferences for personalized recommendation","authors":"Jungmin Hwang , Hakyeon Lee","doi":"10.1016/j.elerap.2025.101527","DOIUrl":"10.1016/j.elerap.2025.101527","url":null,"abstract":"<div><div>Individual preferences change over time. While sequential recommenders have gained attention for accommodating changing user preferences, they have struggled to identify users’ preferences at a granular, component-wise level. This paper introduces a novel approach called preference tracing, inspired by the concept of knowledge tracing, originally developed in the educational domain. Knowledge tracing dynamically estimates a student’s knowledge state through interactions with question–answer pairs and knowledge components, predicting the likelihood of correctly answering an exercise based on the estimated knowledge state. Similarly, preference tracing continuously estimates a user's preference state as they engage with content over time, predicting whether a user will enjoy a specific movie based on the estimated preference state. Our empirical evaluations demonstrate that Bayesian knowledge tracing (BKT)-based preference tracing not only delivers comparable predictive performance but also effectively captures users’ preference states at a component-wise level. Moreover, deep learning-based knowledge tracing (DLKT)-based preference tracing, which operates without predefined movie components, outperforms recent deep learning-based recommendation models, unveiling its potential to provide more accurate and nuanced recommendations.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"73 ","pages":"Article 101527"},"PeriodicalIF":5.9,"publicationDate":"2025-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144606170","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}

引用次数: 0

IntentRec: Incorporating latent user intent via contrastive alignment for sequential recommendation IntentRec：通过顺序推荐的对比校准来结合潜在的用户意图

IF 5.9 3区管理学

Electronic Commerce Research and Applications Pub Date : 2025-06-24 DOI: 10.1016/j.elerap.2025.101522

Seonjin Hwang , Younghoon Lee

{"title":"IntentRec: Incorporating latent user intent via contrastive alignment for sequential recommendation","authors":"Seonjin Hwang , Younghoon Lee","doi":"10.1016/j.elerap.2025.101522","DOIUrl":"10.1016/j.elerap.2025.101522","url":null,"abstract":"<div><div>Predicting the next item a user will interact with is a core task in sequential recommendation (SR). Traditional approaches predominantly focus on modeling patterns in item purchase sequences, yet often fall short in uncovering the underlying motivations behind user behavior. To overcome this limitation, we introduce IntentRec, a novel SR framework designed to incorporate latent user intent signals extracted from user-written reviews. Unlike conventional models that treat item sequences in isolation, IntentRec bridges the semantic gap between review content and behavioral data by aligning their representations in a shared embedding space through contrastive learning. Review sequences chronologically ordered text reflecting users’ thoughts serve as a rich source of intent, which is fused into the item sequence representation during training. To ensure practicality in real-time recommendation scenarios, our method excludes review inputs at inference time, acknowledging that reviews naturally occur after item interactions. IntentRec employs BERT, a pre-trained language model, to extract nuanced user intent from textual reviews, and introduces a cross-attention-enhanced contrastive loss to tightly couple review-derived signals with item-based preferences. Extensive experiments conducted on four widely-used SR benchmarks demonstrate that IntentRec consistently outperforms eight state-of-the-art baselines. Further ablation studies confirm the crucial role of review-based user intent in improving sequential recommendation accuracy.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"73 ","pages":"Article 101522"},"PeriodicalIF":5.9,"publicationDate":"2025-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144480940","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}

引用次数: 0

“Domino effects on eWOM?” understanding consumers’ dynamic perceptions of online travel reviews and perceived travel risk: A three-stage longitudinal approach “对eom的多米诺效应？”了解消费者对在线旅游评论的动态看法和感知的旅游风险：一个三阶段纵向方法

IF 5.9 3区管理学

Electronic Commerce Research and Applications Pub Date : 2025-06-21 DOI: 10.1016/j.elerap.2025.101526

Tao Sun , Junjiao Zhang , Han Zhou

{"title":"“Domino effects on eWOM?” understanding consumers’ dynamic perceptions of online travel reviews and perceived travel risk: A three-stage longitudinal approach","authors":"Tao Sun , Junjiao Zhang , Han Zhou","doi":"10.1016/j.elerap.2025.101526","DOIUrl":"10.1016/j.elerap.2025.101526","url":null,"abstract":"<div><div>Although the impact of the COVID-19 pandemic is gradually diminishing, its influence still persists through people’s experience of travel consumption, including travel risk perception and cautious information processing modes of online travel reviews (OTRs). Since the onset of COVID-19, literature has witnessed an upsurge in illuminating tourists’ intro-pandemic risk perceptions and information behaviors. However, from an evolutionary perspective, a whole spectrum to trace and compare the variations in tourist risk perception and OTR evaluation patterns over time remains unclear. Spanning three investigations pre-, during, and post-pandemic (in 2019, 2020, and 2023), results generally confirm that people’s perception of travel risk has undergone an inverted-U-shaped change, yet perceived equipment risk still maintains at a high level. Additionally, drawing upon the information adoption model (IAM), the results indicate that individuals increasingly consider the argument quality cues (informativeness, persuasiveness) and source credibility cues (expertise, trustworthiness, homophily) of online travel reviews as important over time. The dynamic relationships among different attributes of online travel reviews, perceived information usefulness, and perceived travel risk were also illuminated. Theoretically, findings of this study enriched our understanding of the dynamic role of IAM elements in predicting information usefulness and perceived travel risk in different phases of a public health crisis context. Practically, this study not only provides guidelines on post-pandemic risk management for tourism and hospitality managers, but also gives specific advice for travel websites to best optimize their marketing communication strategies through online reviews in alliance with different risk communication contexts.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"73 ","pages":"Article 101526"},"PeriodicalIF":5.9,"publicationDate":"2025-06-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144365374","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}

引用次数: 0

Personalize it, no return: Nudging online consumers towards product personalization that makes the product non-returnable with herd instinct and regret nudges 个性化，无回报：通过从众本能和后悔推动，将在线消费者推向产品个性化，使产品不可退货

IF 5.9 3区管理学

Electronic Commerce Research and Applications Pub Date : 2025-06-13 DOI: 10.1016/j.elerap.2025.101525

Changyuan Feng , Francisco J. Martínez-López , Yangchun Li , Jordi Campo-Fernandez

{"title":"Personalize it, no return: Nudging online consumers towards product personalization that makes the product non-returnable with herd instinct and regret nudges","authors":"Changyuan Feng , Francisco J. Martínez-López , Yangchun Li , Jordi Campo-Fernandez","doi":"10.1016/j.elerap.2025.101525","DOIUrl":"10.1016/j.elerap.2025.101525","url":null,"abstract":"<div><div>Massive ecommerce returns incur considerable return costs for online sellers, erode their competitiveness, burden their returns systems, and damage the natural environment. Reducing ecommerce returns can mitigate these negative consequences. Since most online sellers adopt a no-return policy for personalized products, inducing consumers to personalize more products should be an effective way for these sellers to reduce ecommerce returns. This article focuses on how online sellers use a herd instinct nudge and a regret nudge to induce consumers to use a product personalization service to reduce ecommerce returns. We also studied the effects of the nudges on several pivotal consumer perceptions and affects. A two-factor (a herd instinct nudge vs. no herd instinct nudge; a regret nudge vs. no regret nudge), between-subject experiment was conducted. This research revealed that both using a herd instinct nudge and using a regret nudge can lead to more consumer product personalization behaviors. Both nudges can make consumers perceive the service more valuable. Compared to a regret nudge, a herd instinct nudge should be a more superior method to induce consumer to use the service because it can increase consumer satisfaction with the seller but did not have a significant influence on consumer perceived threat to decision-making freedom. No interaction effect was found between the two nudges.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"73 ","pages":"Article 101525"},"PeriodicalIF":5.9,"publicationDate":"2025-06-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144579681","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}

引用次数: 0

The coherent two-phased process from sold online to redemption offline on an online daily-deal platform 在线团购平台从线上销售到线下赎回的连贯两阶段过程

IF 5.9 3区管理学

Electronic Commerce Research and Applications Pub Date : 2025-06-13 DOI: 10.1016/j.elerap.2025.101523

Yingxin Song , Yezheng Liu , Xiayu Chen , Muhammet Deveci , Carol Xiaojuan Ou , Lingfei Li , Weizhong Wang

{"title":"The coherent two-phased process from sold online to redemption offline on an online daily-deal platform","authors":"Yingxin Song , Yezheng Liu , Xiayu Chen , Muhammet Deveci , Carol Xiaojuan Ou , Lingfei Li , Weizhong Wang","doi":"10.1016/j.elerap.2025.101523","DOIUrl":"10.1016/j.elerap.2025.101523","url":null,"abstract":"<div><div>Daily-deal platforms closely cooperate with local retailers when issuing daily-deal coupons to profit from selling coupons online and redeeming them offline. However, most research on daily-deal business has only focused on online sales or the offline redemption process. We investigate the coherent two-phased process from selling coupons online to redeeming them offline, grounded in the lens of social judgment theory, to capture the full picture of the daily-deal business. By tracking the sales and redemption of 11,290 deals over a 13-month period on an online daily-deal platform and conducting various data analyses, we find that reputation and price curvilinearly affect the sold online of daily-deal coupons, which consequently positively affects coupon redemption offline. More specifically, the <em>U</em> test empirically indicates that the extreme point of the inverted U-shaped effect of reputation score is 86.0035 within the range [49.7353, 92.7551]. And the extreme point to price demonstrates a U-shaped effect is 399.6082 within the range [4.7060, 829.3651]. We further classify retailers’ daily deals into consumption on a group or individual level. Empirical data demonstrate that the inverted U-shaped effects of reputation and the U-shaped effects of price are weakened by group consumption. Furthermore, we investigate the moderating role of agglomeration on the relationship between daily-deal coupons sold online and redemption offline of daily-deal coupons. We also discussed the theoretical and practical implications.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"73 ","pages":"Article 101523"},"PeriodicalIF":5.9,"publicationDate":"2025-06-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144306573","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}

引用次数: 0

The influence of virtual streamer on purchase intention: The moderated mediating effect of message strategy and live-streaming environment 虚拟主播对购买意愿的影响：消息策略和直播环境的调节中介作用

IF 5.9 3区管理学

Electronic Commerce Research and Applications Pub Date : 2025-06-12 DOI: 10.1016/j.elerap.2025.101520

Xueying Wang, Yuexian Zhang

{"title":"The influence of virtual streamer on purchase intention: The moderated mediating effect of message strategy and live-streaming environment","authors":"Xueying Wang, Yuexian Zhang","doi":"10.1016/j.elerap.2025.101520","DOIUrl":"10.1016/j.elerap.2025.101520","url":null,"abstract":"<div><div>In the realm of live-streaming, virtual streamers represent a significant area of industry expansion. Despite the widespread adoption of both human-backed and AI-backed virtual streamers in marketing campaigns, the differential effects of these two types remain underexplored. Through three experimental studies, this research systematically examines how virtual streamer types (AI-backed vs. human-backed) influence purchase intention, while elucidating the underlying mechanisms and boundary conditions. The findings demonstrate three key insights: First, human-backed virtual streamers exert significantly stronger impacts on purchase intention compared to their AI counterparts, with perceived usefulness serving as the critical mediator. Second, a two-sided message strategy outperforms positive unilateral messaging in amplifying virtual streamers’ effectiveness via enhanced perceived usefulness. Third, the live-streaming environment moderates this mechanism differentially: human-backed streamers prove more effective in real-life environments, whereas AI-backed streamers show superior performance in virtual environments. Both effects operate through the pathway of perceived usefulness. This study advances theoretical understanding of virtual streamer efficacy while providing actionable guidelines for businesses to optimize streamer selection strategies.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"73 ","pages":"Article 101520"},"PeriodicalIF":5.9,"publicationDate":"2025-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144306605","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}

引用次数: 0

Large language model meets chaos: A new deep learning model for fake review detection 大语言模型遇上混沌：一种新的虚假评论检测深度学习模型

IF 5.9 3区管理学

Electronic Commerce Research and Applications Pub Date : 2025-06-10 DOI: 10.1016/j.elerap.2025.101521

Yu Fan , Haizhou Fan , Fuqian Zhang , Zhenhua Wang

{"title":"Large language model meets chaos: A new deep learning model for fake review detection","authors":"Yu Fan , Haizhou Fan , Fuqian Zhang , Zhenhua Wang","doi":"10.1016/j.elerap.2025.101521","DOIUrl":"10.1016/j.elerap.2025.101521","url":null,"abstract":"<div><div>Detecting fake online reviews is crucial for the e-commerce ecosystem. However, existing studies often fail to mine the intrinsic attributes of reviews, which limits detection performance. In this paper, we introduce a novel fake review detection model, LLMChaos, which investigates reviews from the perspectives of large language models (LLMs) and chaos theory. Specifically, we first propose a method that blends energy selection with LLMs to generate review time series. Second, we construct a space mapping mechanism with multiple chaotic attributes, embracing the intrinsic attributes of reviews. Finally, we design a hierarchical learning network that trains in deep contrastive learning across LLM layers, chaotic attribute layers, and Transformer layers. Extensive experiments demonstrate that LLMChaos is robust and state-of-the-art. For instance, on the Hotel dataset, LLMChaos achieves 94.78% F1, outperforming recent models by 1.42%-19.78%; on the Amazon dataset, LLMChaos achieves 93.15% F1, surpassing recent models by 1.45%-18.39%. Moreover, we contribute novel discoveries, for example, chaotic behaviors of reviews generally exhibit bounded ranges: Lyapunov exponent (0–0.0125), Correlation dimension (0.25–0.5), Kolmogorov entropy (0.75–0.85), Fractal spectrum (0–1.5), and Recurrence rate (0.005–0.015); real and fake reviews display distinct chaotic distributions.</div></div>","PeriodicalId":50541,"journal":{"name":"Electronic Commerce Research and Applications","volume":"73 ","pages":"Article 101521"},"PeriodicalIF":5.9,"publicationDate":"2025-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144270313","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}

引用次数: 0