18 September 2025, Volume 46 Issue 5

    

• Select all
  |

• Select
  Research on the Logical Approach and Technical Pathway of Multi-Source and Multi-Modal Data Fusion in Intelligence Analysis

  Xia Yikun, Liu Bowen, Tian Cong

  Information and Documentation Services. 2025, 46(5): 5-15. https://doi.org/10.12154/j.qbzlgz.2025.05.001

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  [Purpose/significance] As the frontier of data utilization, the intelligence analysis still faces challenges such as decoupling of thinking and technology, the emergence of marginal effects, and poor adaptation between tasks and technologies in the fusion of multi-source and multi-modal data. It is urgent to explore the logical thinking and technical path for data fusion and utilization under the guidance of intelligence thinking and the traction of intelligence tasks. [Method/process] This paper clarifies the ubiquity, high dimensionality, dynamicity, complementarity and re⁃dundancy characteristics of multi-source and multi-modal data, and explains the multi-source and multi-modal data fusion for intelligence analysis from the perspective of information chain and information theory; based on the principle of "data is useful, available and utility", the underlying logic is constructed to form the basic idea of organic coordina⁃tion of intelligence tasks, data characteristics and system effectiveness; by embedding digital intelligence technology in⁃to the intelligence analysis process, a three-stage progressive multi-source and multi-modal data fusion technical path of "data resource preparation, multi-dimensional feature alignment, smart intelligence generation" is proposed. [Re⁃sult/conclusion] The research results provide strategic references and methodological guidelines for the fusion of multi-source and multi-modal data to empower intelligence practice, which is helpful to promote the deep mining of da⁃ta value and the digital-intelligent transformation of intelligence work.
  
• Select
  Questioning Intelligence Forecasting and Exploring Feasible Pathways

  Peng Zhihui

  Information and Documentation Services. 2025, 46(5): 16-23. https://doi.org/10.12154/j.qbzlgz.2025.05.002

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  [Purpose/significance] This paper critically questions and reflects on information forecasting, aiming to deepen the understanding of its characteristics and limitations and promote its sound development. [Method/process]The paper first analyzes the internal logic and theoretical presuppositions of information forecasting, examines its theo⁃retical foundations and questions the rationality of these foundations, and finally proposes the feasible approaches of in⁃formation forecasting. [Result/conclusion] The internal logic of information forecasting is "predicting events based on historical patterns", i.e., inferring the unknown from the known and speculating about the future based on the past and present states of things. This internal logic relies on a fundamental theoretical presupposition: the development of things is governed by social laws, which can be identified through analyzing information about their past and present states, and thus the future of things can be inferred from these laws. However, the theoretical foundations supporting the
  internal logic and theoretical presuppositions of information forecasting—namely, the theory of future causality and the principle of continuity—both contain obvious logical flaws and lack internal logical consistency. Therefore, information forecasting should be conducted with foreknowledge as its purpose, indicator monitoring as its method, and information assessment as its core content.
  
• Select
  Analysis of Generative Artificial Intelligence′s Information Resource Management Functions:Causes, Challenges, and Responses

  Sun Yazhou, Li Xiaosong, Tang Shanhong, Hua Juan

  Information and Documentation Services. 2025, 46(5): 24-34. https://doi.org/10.12154/j.qbzlgz.2025.05.003

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  [Purpose/significance] In the context of digital information, the social function of information resource man⁃agement has changed from information resource provider to knowledge service innovator, social problem solver and cul⁃tural ecology builder, and the user information acquisition paradigm has changed from "search - screening" to "ques⁃tion - generation". Generative artificial intelligence represented by DeepSeek, as a new channel for information acqui⁃sition, is gradually regarded by users as an information service model similar to information resource management.[Method/process] Focusing on the phenomenon of the formation and exertion of the information resource management function of generative artificial intelligence, exploring its causes from multiple dimensions such as technology, society and cognition, analyzing the social benefits of generative artificial intelligence as an information channel under techno⁃logical advantages, and clarifying the resulting technology intermediary cognition and its alienation challenges. [Result/
  conclusion] Explored the construction strategies of the human-machine collaborative cognitive ecosystem, provided references for solving the challenges brought by generative artificial intelligence in performing the function of informa⁃tion resource management and developing digital intelligence-driven information services.
  
• Select
  Research on the Influencing Factors of Learning Effect in Learning-related Search in Human-AI Interaction Contexts

  Xu Dan, Song Xiaoxuan, Gong Hong

  Information and Documentation Services. 2025, 46(5): 35-44. https://doi.org/10.12154/j.qbzlgz.2025.05.004

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  [Purpose/significance] The human-AI search context has significantly reshaped behavioral interactions and learning effect in learning-related search. This study aims to provide a comprehensive evaluation of learning effect and to examine how search tools and task types jointly influence learning outcomes. [Method/process] An experimen⁃tal method was adopted to design three types of learning-related search tasks: receptive, critical, and creative. Two kinds of search tools, traditional search engines and Generative AI (GenAI), were provided. Learning effect data were collected from users who completed different tasks with these tools. The analysis of learning effect was conducted from two perspectives: subjective learning experiences and objective learning outcomes.[Result/conclusion] Users′ learn⁃ing experiences were mainly influenced by task types, while the specific effect of the search tool used remained un⁃clear. In terms of learning outcomes, traditional search engines outperformed GenAI in receptive and critical tasks, par⁃ticularly in the number, coverage, and uniqueness of knowledge points as well as the depth of knowledge facets. For cre⁃ative tasks, however, both tools provided relatively limited support, with GenAI showing a slight advantage.
  
• Select
  Research on the Moderating Effect of Attention Allocation Concentration in the Adoption of User Information by Government Open Data Users

  Zhao Xinyue, Zhao Jing, Zhou Xia, Wang Yunjiao

  Information and Documentation Services. 2025, 46(5): 45-56. https://doi.org/10.12154/j.qbzlgz.2025.05.005

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  [Purpose/significance] Researching the adoption of government open data user information is conducive to enhancing the readability and accessibility of data, triggering users to focus their attention, and providing theoretical references for in-depth data value mining. [Method/process] Using the SOR (Stimulus-Organism-Response) frame⁃work as the analytical structure, and integrating the theories of narrative transportation, perceived value, trust, and at⁃tention-based view, this study analyzes narrative elements such as narrative content, structure, and data presentation methods, explores the moderating effect of attention allocation concentration, constructs a theoretical hypothesis model of factors influencing the willingness of government open data user information adoption, and empirically verifies the model through a survey of 572 government open data users. [Result/conclusion] The study shows that the narrative content, structure, and presentation methods of government open data positively affect cognitive trust and perceived val⁃ue, and perceived value positively affects cognitive trust. Both cognitive trust and perceived value positively affect user
  information adoption willingness and play a mediating role in the process of influencing adoption willingness through the stimulus dimension. Attention allocation concentration positively moderates the impact of cognitive trust and per⁃ceived value on user information adoption willingness.
  
• Select
  Research on the Identification and Formation Path of High-Value Interactions between Cross-Platform Bloggers and Their Audiences

  Yan Weiwei, Xiong Xinyu

  Information and Documentation Services. 2025, 46(5): 57-67. https://doi.org/10.12154/j.qbzlgz.2025.05.006

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  [Purpose/significance] With the diversification of online platforms, users are increasingly inclined to carry out information exchange activities across multiple platforms, therefore, focusing on cross-platform users, exploring their interaction types, identifying high-value interactions and analyzing the path of their formation are of great signifi⁃cance in understanding cross-platform users' information exchange behaviors. [Method/process] By aligning bloggers who have cross-platform interactions on representative social sharing platforms and short-video platforms, and obtain⁃ing all comments under the content they posted throughout the year of 2023, we implement the division of interaction types based on the K-means method to identify high-value interactions and analyze their formation paths. [Result/con⁃clusion] Two-way interactions between cross-platform bloggers and audiences can be divided into two categories: ac⁃tive interactions and loose interactions, both of which appear more often on the social sharing platforms. Active interac⁃tions are more frequent, involve a wider range of topics, and have a more intimate relationship than loose interactions,
  which helps both parties to form social capital, and can be regarded as an online extension of interpersonal relation⁃ships, and is a high-value interaction. The formation of such high-value interaction needs to go through three stages:one-way attention, surface contact and mutual involvement, and there are three main formation paths: approval-contri⁃bution type, demand-exchange type and care-sharing type, among which approval-contribution type is the type that appears most frequently on social sharing platforms, demand-exchange type is the type that occurs the most on shortvideo platforms, and care-sharing type is the type that occurs less often on both platforms.
  
• Select
  Identification and Analysis of Key Influencing Factors of Users′ Discontinuance Behavior in Generative AI

  Mi Qinze, Qiu Junping, Xu Zhongyang, Hu Bo, Zhou Tao

  Information and Documentation Services. 2025, 46(5): 68-77. https://doi.org/10.12154/j.qbzlgz.2025.05.007

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  [Purpose/significance] After experiencing unprecedented user growth, generative AI has encountered a sig⁃nificant decline in user scale for the first time. Identifying and analyzing the key factors influencing generative AI user discontinuance behavioris of great significance for optimizing user experience, enhancing user stickiness, and promot⁃ing the healthy development of the generative AI industry. [Method/process] This study used a combination of litera⁃ture survey and Delphi methods to systematically and comprehensively distill the influencing factors on discontinuance behavior among generative AI users, and identified and analyzed the key influencing factors using the DEMATEL meth⁃od. [Result/conclusion] The research findings indicate that: the discontinuance behavior of generative AI users is in⁃fluenced by 20 influencing factors in four dimensions: user factors, information factors, platform factors, and environ⁃mental factors; algorithmic literacy, degree of intelligence, degree of anthropomorphism, competitive environment and other 10 influencing factors are all key influencing factors of generative AI users′ discontinuance behavior; generative AI platform managers should reduce the occurrence of users′discontinuance behavior by deeply understanding users′ needs, improving users′ algorithmic literacy, establishing a social interaction platform and other aspects.
  
• Select
  From Digital Disconnection to Knowledge Disconnection: Research on Behavioral Response Mechanisms and Practical Paradoxes Based on the SOR Model

  Yao Wei, Liu Cui, Wang Lichen, Han Yujia, Shi Mengru, Wang Shaonan

  Information and Documentation Services. 2025, 46(5): 78-88. https://doi.org/10.12154/j.qbzlgz.2025.05.008

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  [Purpose/significance] The phenomenon of involution alters the individual's cognition of competitive pres⁃sure, generating irrational knowledge learning behaviors. Based on the SOR model, it is significant to reveal the intrin⁃sic mechanism of knowledge disconnection and identify the paradoxical relationship between knowledge disconnection and connection, so as to understand knowledge disconnection and intervene dynamically in order to promote the sus⁃tainable development of knowledge subjects and society. [Method/process] This study proposes the concept of knowl⁃edge disconnection based on digital disconnection related research, analyzes the triggering basis, reaction process and implementation strategy of knowledge disconnection, and explores the practical paradoxes of knowledge disconnection and builds a framework for solving the paradoxes on this basis. [Result/conclusion] There are three kinds of behavioral responses, namely filtered, defensive and ambiguous knowledge disconnection, to regulate the mental state and inner perception under multi-source conditioned stimuli. There are value paradoxes, emotional paradoxes and cognitive para⁃doxes in the process of knowledge disconnection practice, which are manifested in the opposing choices of knowledge
  disconnection and connection, triggering individuals' paradoxical thinking. This study also constructs a paradoxes reso⁃lution framework that includes three processes: inclusive acceptance, harmonious adaptation, and dynamic synergy, in order to promote the realization of the dynamic balance and virtuous cycle of knowledge disconnection and connection.
  
• Select
  From Evolution to Collaboration: Exploring the Path to Enhancing Efficiency in Government Data Governance Policies within the GraphRAG Framework

  Zhu Xiaofeng, Qin Deming

  Information and Documentation Services. 2025, 46(5): 89-101. https://doi.org/10.12154/j.qbzlgz.2025.05.009

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  [Purpose/significance] In the context of data-driven governance, solving the problem of fragmented govern⁃ment data governance policies and achieving overall policy effectiveness improvement have become key breakthroughs in unleashing the value of data elements and building a modern governance system. [Method/process] This article is based on the third-order research paradigm of "dynamic evolution analysis→ collaborative linkage mining → efficien⁃cy improvement path extraction". On the basis of constructing the theoretical framework of government data governance policy efficiency improvement，relying on the GraphRAG framework and integrating the deep semantic reasoning abili⁃ty of the big language model with the multidimensional relationship mapping technology of the knowledge graph, this ar⁃ticle analyzes the evolution process and evolution law of China's government data governance policy. Furthermore, it an⁃alyzes the collaboration of policy subjects and the linkage of policy content elements, explores the collaborative difficul⁃ties of China's government data governance policy, and finally proposes efficiency improvement path suggestions, pro⁃viding a practical direction of "evolution interpretation collaborative optimization efficiency improvement" for optimiz⁃ing government data governance policy. [Result/conclusion] China's government data governance policies have evolved from "block segmentation" to "multi-dimensional collaborative ecology", from "single point breakthrough" to "system empowerment", and from "passive response" to "value consciousness" in cognition; there are collaborative challenges such as "cellular isolation" and "breadth intensity imbalance"; it is necessary to build a cross domain collab⁃orative governance ecosystem, improve the dynamic adaptation mechanism of policy content, and enhance the overall synergy of multidimensional policies in order to improve the overall effectiveness of policies.
  
• Select
  Research on the Construction of Tendency Dictionary for Chinese Policy Texts from the Perspective of Sentiment Analysis

  Wang Yu, Wu Bin

  Information and Documentation Services. 2025, 46(5): 102-112. https://doi.org/10.12154/j.qbzlgz.2025.05.010

  Abstract ( ) Download PDF ( )   Knowledge map   Save

  [Purpose/significance] It is of great value to grasp the concept of government behavior and governance ac⁃curately, based on the dictionary method to construct a fine grain analysis of Chinese policy texts suitable for the open data environment. [Method/process] The words that reflect emotional intensity in policy texts are defined as tendency words, and their structural characteristics and semantic relevance are used to construct a tendency dictionary. First, the seed words are extracted according to the interpretation opinions of domain experts and combined with the point mutual information algorithm to expand the dictionary online. Secondly, based on the theory of formal concept analysis, the top⁃ic connotation of policy texts is defined and quantified, the hierarchical relationship between policy topics is mapped to the semantic relationship between words, and the synonymous tendency words with topic similarity are screened. Final⁃ly, the credibility and validity methods are used for empirical testing. [Result/conclusion] The propensity dictionary
  has high accuracy and recall rate in the task of policy text emotion recognition, which is suitable for large-scale fine grain policy text analysis, and provides a reliable and novel quantitative tool for policy research and decision-making.