2026 4 works
Analysis of author feedback behavior in different peer review decision-making scenarios： taking the F1000research platform as an example
Zhaoping Yan, Bo Xiang, Jin Shi†(†corresponding author)
Chinese Journal of Scientific and Technical Periodicals 2026 Journal
Purposes This study systematically analyzes author feedback behavior in peer review interactions， providing practical references for authors to strategically respond and deepening our understanding of the mechanisms of peer review interaction. Methods Taking the open peer review platform F1000Research as the research object， a quantitative analysis was conducted on author feedback behavior under different peer review decision-making scenarios. Machine learning and statistical analysis methods were used to conduct in-depth analysis of author feedback content from three aspects： linguistic features， text content， and the focus of author feedback. A logistic regression model is constructed to further analyze the potential mechanisms by which author feedback behavior influences reviewer decision-making. Findings The results indicate that positive revisions have a significant potential to reverse reviewer attitudes. The author’s sentiment and politeness in the response have a significant impact on the reviewers’ attitude change. In author responses that reverse reviewers’ decision， improvements related to research design and research subjects emerge as the most influential factors in persuading reviewers. Conclusions Analyzing author feedback behavior reveals the relationship between feedback characteristics and peer review decisions under different review decision-making scenarios. This not only provides authors with effective feedback strategies but also offers practical guidance for journals to optimize their review processes and improve review efficiency.
The technological and societal co-evolution landscape of AI medical conversational agents: An LLM-enhanced quantitative textual review
Bo Xiang, Zhaoping Yan, Hongcheng Wei, Jin Shi†(†corresponding author)
Expert Systems With Applications 2026 Journal
As an interdisciplinary product of artificial intelligence (AI), medical informatics, and human-computer interaction, AI medical conversational agents integrate cutting-edge technologies such as large language models (LLMs) and intelligent agents. Consequently, their application has given rise to numerous complex socio-ethical challenges. Grounded in sociotechnical system theory, this research constructs a mixed bibliometric framework integrating LLM, topic modeling, and social network analysis, which is used to explore the co-evolution landscape of technology and society in the field of AI medical conversational agents. In particular, a total of 4,649 domain-relevant bibliographic records are collected from the Web of Science and Engineering Village. Then, a hybrid architecture combining LLM-based agents and human verification is applied to data cleaning and alignment. Leveraging the language comprehension capabilities of LLMs, relevant literature is mapped into societal and technological dimensions, aligning with the sociotechnical system theory. The evaluation demonstrates that the few-shot approach yields better performance than the zero-shot approach across accuracy, precision, recall, and F1-score. Structural topic model is employed to construct a knowledge map comprising 5 main domains and 14 key research themes in the field of AI medical conversational agents. Based on the social and technical categorization of the literature, these topics are delineated into technological, societal, and neutral types. The topic correlation network reveals dense intra-category connections, whereas direct linkages between societal and technological topics are sparse, relying primarily on intermediate topics for bridging. Combining keyword-based analysis with the disciplinary classification of literature, this study unveils the intrinsic mechanism of the socio-technological co-evolution landscape. Notably, themes revolving around interactive experience and clinical utility serve as the primary bridges connecting social and technical subsystems. Overall, this study proposes a methodological framework integrating LLMs for bibliometric analysis, providing fresh insights into the critical domain of AI medical conversational agents.
Emotional comfort or technological delusion? An exploration of public perceptions and attitudes toward AI companions based on video social medias
Hongcheng Wei, Bo Xiang, Xihun Zheng, Jianlin Yang†(†corresponding author)
International Journal of Human-Computer Interaction 2026 Journal
AI companions, driven by artificial intelligence, are increasingly embedded in users’ daily lives via anthropomorphic and interactive conversational experiences. Although these systems can offer emotional support, they also raise concerns regarding algorithmic bias or data security. A comprehensive understanding of public perceptions is therefore essential for promoting healthy human-AI relationships. This paper investigates public perceptions of AI companions using 613 videos and 33,133 comments from two video platforms. Combining text mining and large language model-based sentiment analysis, it examines public topics and sentiment patterns regarding AI companions from multiple dimensions. Public attention surged after generative AI’s rise in 2023, with a pronounced “East-West divide” in user engagement. Discussions cover society (role substitution, family) and technology (companionship, anthropomorphism), with notable differences in the topics between men and women, and overall sentiment is positive. Moreover, personality traits also influence attitudes. By comparing these results with existing studies, it highlights how Chinese cultural contexts shape public perceptions. Drawing on diverse user-generated data from video platforms, the study offers valuable insights for the better design, marketing, and policymaking of AI companions.
A study on evaluating the impact of dissertations based on the integration of academic and social dimension characteristics
Zhaoping Yan, Bo Xiang, Jin Shi†(†corresponding author)
Information Science 2026 Journal
Existing dissertation impact evaluation indexes have limitations, which make it difficult to comprehensively and accurately assess the contribution of research achievements, to solve this problem, this paper aims to construct a dissertation impact evaluation model that integrates academic and social dimensions This paper constructs an impact evaluation model from both academic and social dimensions. In the academic dimension, citation is weighted by integrating influential factors such as citation position, citation sentiment, citation relevance and temporal heterogeneity to construct the academic impact index. In the social dimension, the indicators are selected from five aspects: use, access, mention, social media and citation to construct the social impact evaluation framework The empirical analysis shows that the paper impact evaluation model constructed in this paper can more comprehensively and accurately assess the impact of papers and identify high-impact papers. The sensitivity analysis of the weights and the comparative analysis found that the model has good robustness and effectiveness. This paper proposes a new method for evaluating academic papers that integrates academic impact and social impact, which can reflect the dual impact of papers in the academic and social fields, thereby more effectively identifying high-impact papers.Tracking technology evolution paths is crucial for understanding innovation dynamics. However, current studies frequently only use text-based methods, which limit their capacity to capture the structural linkages found in technology knowledge. To address this gap, this study proposes an integrated framework that combines text embedding and network embedding to analyze technology evolution paths. Using a dataset of 35983 patents in the battery electric vehicle (BEV) domain, we extract high-dimensional semantic features from patent texts while simultaneously capturing structural relationships between patents by constructing a heterogeneous information network. By fusing these two embeddings, we construct technology evolution paths based on cosine similarity measures. The proposed method effectively reveals knowledge diffusion patterns and emerging technology trends. In addition, through pair sample t-test and topic coherence analysis, we found that the proposed method identifies more comprehensive technology evolution trajectories.
2025 4 works
A hybrid embedding method for identifying technology evolution paths of patents: The case on battery electric vehicle industry
Zhaoping Yan, Bo Xiang, Jin Shi†(†corresponding author)
IEEE Transactions on Engineering Management 2025 Journal
Tracking technology evolution paths is crucial for understanding innovation dynamics. However, current studies frequently only use text-based methods, which limit their capacity to capture the structural linkages found in technology knowledge. To address this gap, this study proposes an integrated framework that combines text embedding and network embedding to analyze technology evolution paths. Using a dataset of 35983 patents in the battery electric vehicle (BEV) domain, we extract high-dimensional semantic features from patent texts while simultaneously capturing structural relationships between patents by constructing a heterogeneous information network. By fusing these two embeddings, we construct technology evolution paths based on cosine similarity measures. The proposed method effectively reveals knowledge diffusion patterns and emerging technology trends. In addition, through pair sample t-test and topic coherence analysis, we found that the proposed method identifies more comprehensive technology evolution trajectories.
Detecting technology opportunities appropriate for enterprise R&D: The synthesis analysis of industrial technical windows and enterprise competition relations
Bo Xiang, Zhuoya Pan, Dejian Yu†, Wenjin Zuo(†corresponding author)
Technology in Society 2025 Journal
Technological opportunities (TOs) are the potential and set of possibilities for technology advances in a given industry. When enterprises are able to catch and adapt to them in a timely manner, they can grab market share from competitors who have failed to adapt to these challenges. However, when there exist large gaps between enterprises and their competitors, it should be carefully evaluated whether enterprise-specific TOs are worth exploring. Moreover, faced with diversified competitive relations, enterprises also need to formulate differentiated research and development (R&D) strategies for different TOs. To address these research gaps, this paper argues for the theoretical concepts of technical windows (TWs), emerging technologies (ETs), and TOs, and proposes a three-stage framework to detect enterprise-specific TOs.
Customized integrated decision model for CBEC enterprise credit evaluation: The fusion of multi-source features and machine learning
Dejian Yu, Bo Xiang†(†corresponding author)
Electronic Markets 2025 Journal
The cross-border e-commerce (CBEC) industry plays a crucial role in the transformation of foreign trade and the upgrading of innovative development, driven by information technology and international trade policies. However, the distinctive operational pattern of CBEC enterprises necessitates the customization of the corporate credit evaluation framework to their specific features, which is absent in the existing studies. This paper proposes an integrated decision framework that incorporates multi-source features and machine learning algorithms to achieve customized credit evaluation for CBEC enterprises.
Theoretical framework on synergistic mechanism and extraction strategy of scientific and technological knowledge driven by data-intelligence
Bo Xiang, Zhaoping Yan, Zhuoya Pan, Dejian Yu, Jin Shi†(†corresponding author)
Journal of Information Resources Management 2025 Journal
It is the important issue to promote the value realization of data elements through clarifying the multiple stakeholders and multi-source heterogeneous data of scientific and technological knowledge, and their synergistic mechanism, as well as delineating the knowledge extraction paths empowered with big data and intelligent technology. Based on the process of data-to-wisdom derivation in the DIKW chain, a synergistic framework of scientific and technological knowledge involving multiple stakeholders, such as universities, research institutes, enterprises, government and the public, has been constructed, covering multi-source data such as papers, patents, products, policies and user-generated content. Then, the front-end and back-end structures of scientific and technological knowledge extraction paths are expanded, as well as the knowledge extraction outcomes and their diverse service scenarios under intelligent strategy combination patterns are explored.
2024 4 works
Identification of the knowledge trajectory of internet of vehicles: From the perspective of main path analysis and topic analysis
Zhaoping Yan, Bo Xiang, Dejian Yu†, Jin Shi†(†corresponding author)
IEEE Internet of Things Journal 2024 Journal
The Internet of Vehicles (IoV), as the cornerstone of intelligent transportation systems, is gradually attracting attention and accumulated a large amount of literature. Therefore, this article employs citation analysis and topic analysis to analyze the research in the IoV field, revealing the knowledge evolution trajectory and development dynamics.
Exploring the evolution and collaboration in two-sided matching: A comprehensive bibliometric and topic modeling analysis
Xiaorong He, Bo Xiang†, Zeshui Xu, Dejian Yu(†corresponding author)
International Journal of Intelligent Computing and Cybernetics 2024 Journal
This study offers a novel and detailed overview of TSM research highlighting significant trends and collaboration patterns within the field. By integrating bibliometric methods with structural topic modeling the study provides unique insights into the evolution of TSM research making it a valuable resource for both academic and professional communities.
An ESTs detection research based on paper entity mapping: Combining scientific text modeling and neural prophet
Dejian Yu, Bo Xiang†(†corresponding author)
Journal of Informetrics 2024 Journal
Existing studies on the detection of emerging scientific topics (ESTs) overemphasize the newness and neglect content innovation of knowledge. Moreover, they also ignore the lag existing in knowledge diffusion. In this paper, we propose a four-stage detection framework for ESTs that maps emerging attributes from paper entities to scientific topics.
Combining text analytics and network path extraction to trace CSR in the social sciences: Intellectual structures and diffusion trajectories
Dejian Yu, Bo Xiang†, Zhuoya Pan(†corresponding author)
Corporate Social Responsibility and Environmental Management 2024 Journal
Corporate social responsibility (CSR) has evolved over time into a mature interdisciplinary scientific field. However, there is a lack of research to explore the disciplinary association patterns and diffusion trajectories of scientific knowledge within this field. This research proposes a knowledge tracing framework that combines text analysis based on scientific texts and main path extraction based on citation networks to fill this gap.
2023 2 works
Discovering topics and trends in the field of Artificial Intelligence Using LDA topic modeling
Dejian Yu, Bo Xiang†(†corresponding author)
Expert Systems with Applications 2023 Journal
Artificial Intelligence (AI) has affected all aspects of social life in recent years. This study reviews 177,204 documents published in 25 journals and 16 conferences in the AI research from 1990 to 2021, and applies the Latent Dirichlet allocation (LDA) model to extract the 40 topics from the abstracts. This study aggregates the results of the LDA model from the perspectives of year, publication source, and country/region. The aggregated result is the topic distribution from different perspectives. Analysis of the aggregated results reveals the research characteristics of different publication sources (countries/regions) in the AI research, and which publication sources (countries/regions) have similar research content. These results provide help for scholars and research institutions to choose research directions, and new entrants to understand the dynamics of the field.
Discovering knowledge map and evolutionary path of HRM and ER Using the STM combined with Word2vec
Dejian Yu, Bo Xiang†(†corresponding author)
International Journal of Manpower 2023 Journal
This work adopts state-of-the-art textual as well as semantic mining techniques to establish a comprehensive knowledge map for HRM and ER research. Furthermore, these results uniquely demonstrate the pluralistic ideological orientation at the social level is gradually integrated into more micro levels, such as enterprises and individuals. These are the contents that were mentioned from previous studies by scholars, but not meticulously verified and interpreted.