Most Download

  • Published in last 1 year
  • In last 2 years
  • In last 3 years
  • All

Please wait a minute...
  • Select all
    |
  • WANG Dakun, HUA Bolin
    Scientific Information Research. 2025, 7(1): 131-140. https://doi.org/10.19809/j.cnki.kjqbyj.2025.01.012
    [Purpose/significance]Identifying and foreseeing emerging technologies, bring technological first-mover advantages to enterprises and governments, and grasp technological development trends in a timely manner. [Method/process]This study uses BERTopic's topic modeling method to obtain domain topic distribution, and merges paper and patent topics based on the cosine similarity of topic vectors to identify emerging topics. [Result/conclusion]Using the BERTopic topic modeling method combined with index evaluation can effectively identify emerging topics and emerging terms.Taking the field of new energy vehicles as an example to carry out empirical research, using two methods: divided verification period and data verification method, 12 of the 16 identified topics passed the verification, which verified the effectiveness of this research method.
  • CHEN Yucheng, LI Yang, LIU Jiangfeng, YANG Fan
    Scientific Information Research. 2024, 6(2): 115.
    [Purpose/significance]Intangible cultural heritage is an important component of human civilization, which is of great significance for protecting and promoting national spirit, enhancing national identity and cohesion. [Method/process]This paper explores how to utilize the advantages of  AIGC, combined with traditional deep learning methods, then construct a comprehensive and efficient map of intangible cultural heritage knowledge. [Result/conclusion]In the classification study of intangible cultural heritage projects, the fine-tuned Baihuan-7B has the best effect, with an macro-F1 value of 0.7688. In the extraction of intangible cultural heritage attribute information, RoBERTa has the best effect, with an F1 value of 0.7085. The 2-Gram BLEU generated by fine-tuning Baihuan-7B is 0.2052. Combining the results of attribute extraction and generation, an efficient and comprehensive knowledge graph is constructed. [Innovation/limitation]This study uses a generative large model to assist in the establishment of knowledge graphs, focus on national-level intangible fruit projects, not induding provinciul-level projects with high reasearch value.
  • DENG Sanhong, HU Haotian, WANG Hao, WANG Dongbo,
    Scientific Information Research. 2021, 3(1): 1.
    [Purpose/significance]With the popularization of digitized ancient books and documents,the use of natural language processing and big data analysis technology to carry out text mining and knowledge discovery on ancient Chinese books has gradually become an important research direction in the field of ancient information processing of digital humanities and an important way to reflect cultural confidence.[Method/process]This article defined the concept of ancient Chinese character automatic processing.Wesorted out the connotation and extension of the ancient Chinese character automatic processing,and grasped the overall research status and development trend of this fieldfrom the three aspects of the field of automatic ancient texts processing and model algorithms,corpus resources and existing tools,knowledge bases and platform system.[Result/conclusion]We conducted a more comprehensive summary of the current research status of ancient Chinese character automatic processing,and analyzed the existing problems and deficiencies.
  • CHEN Yanhong, WANG Huan
    Scientific Information Research. 2024, 6(1): 49.
    [Purpose /significance]In the event of an emergent epidemic, the community is an important channel for vulnerable  groups to obtain emergency information, and understanding the information needs of vulnerable groups will help the community accurately provide information services, promote the emergency information reserves of the residents in the jurisdiction, and enhance their satisfaction. [Method/process]Taking the emergent epidemic as the background,the study combines the important information categories of the epidemic and related literature to investigate the emergency information needs of the disadvantaged groups in the community and classifies each need with the help of KANO questionnaire, in order to determine the priority of the community to optimize the emergency information service. [Result /conclusion]The priorities of emergency information services should be adjusted according to the basic demand,expectation demand,charm demand and irrelevant demand of vulnerable groups. By paying attention to the transformation of emergency needs and broadening the channels of information dissemination, which could guarantee the emergency information needs of vulnerable groups are satisfied.
  • LI Weichao, ZHOU Yi,
    Scientific Information Research. 2023, 5(4): 57-77. https://doi.org/10.19809/j.cnki.kjqbyj.2023.04.005
    [Purpose/significance]This study aims to provide valuable references for future research and governance practices by reviewing the research on the governance of Internet information's content ecosystem in China.[Method/process]This paper collects literature data from the core journal library of CNKI, reveals the macro situation of this topic research by using the bibliometric method, selects the literature that meets the quality standards by using the systematic review method, summarizes relevant studies from four dimensions: governance bodies, governance objects, governance tools and governance mechanisms.[Result/conclusion]The co-governance of multi-subjects  has become the basic consensus of the academic circle. Solving the problem of subjects failure and promoting the coordination of subjects is one of the focuses of attention in recent years. Current research has broken through the information content itself, and all kinds of subjects, technologies and background environment related to the production, dissemination and use of information content have become research objects. Governance tools can be classified into three categories: regulation tools, promotion tools and educational guidance tools. It has become a consensus of the government and most scholars to strengthen the effectiveness of governance tools with the help of digital intelligence technology. Governance mechanism can also be summarized and divided into three stages: before, during and after. In the future, more studies should be carried out from the overall perspective of the governance of internet information's content ecosystem, it is necessary to strengthen the research on the governance of internet positive information's content ecosystem and the algorithm security about internet information's content ecosystem.
  • CAO Shujin, WANG Yaqi, LU Guangxu
    Scientific Information Research. 2022, 4(1): 34. https://doi.org/10.19809/j.cnki.kjqbyj.2022.01.004
    [Purpose/significance]The study focusing on the relationship between the affecting interaction factors and user experience during the use of search engines, is of great significance for the accurate acquisition of information in the subject area, also can improve the efficiency of research, and promote the development of the subject.[Method/process]This research selects Baidu academic search engine as the object.In order to study the interaction factors that affect academic search engine user experience,we build a model of academic search engine interaction factors affecting user experience based on the FMP interaction model, and collect data through questionnaires.[Result/conclusion]We found that: retrieval function interaction and retrieval content interaction have a significant impact on user experience, and content interaction has a greater impact on user experience than functional interaction;operational functions such as classification, sorting, and quick citation of detected results have an impact on retrieval Functional interaction and retrieval content interaction;the design of the retrieval interface has a significant impact on the interaction of the retrieval content;the auxiliary orientation function during the retrieval process has a significant impact on the interaction of the retrieval function.
  • YU Chuanming, WANG Feng, ZHANG Zhengang, KONG Lingge, AN Lu
    Scientific Information Research. 2021, 3(1): 56.
    [Purpose/significance]By combining knowledge graph representation learning and word representation learning,this paper explores the question answering model based on knowledge graph,so as to realize the utilization of structured knowledge with high accuracy and wide coverage in the knowledge graph.[Method/processe]This paper proposes a knowledge-based QA system,a framework integrating knowledge graph representation learning and word representation learning,and uses comparative and empirical research to explore the influence of different representation learning models and network structure on the effect of knowledge-based QA system.Firstly,the knowledge graph representation learning algorithm is used to generate the vector representation of entities and relations. Secondly,the generated entity and relation vectors are used as supervisory signals to train the vector representation of the problem.Finally,the best answer of the matching problem in the knowledge graph is selected by the triple representation generated by the problem.[Result/conclusion]The experimental results show that different representation learning model and network structure have a significant impact on the effect of knowledge-based question answering.Compared with the baseline method,this method can significantly improve the effect of knowledge base question answering.The research plays an important role in promoting the application of deep learning in the research of knowledge-based question answering.
  • ZHANG Hai, FANG Jiping, WANG Dongbo
    Scientific Information Research. 2025, 7(2): 48-57. https://doi.org/10.19809/j.cnki.kjqbyj.2025.02.005
    [Purpose/significance]The deep integration of ancient book digitization and large language models is the future development trend. In order to clarify the influence factors and formation mechanisms of dropout behavior among users of domain ancient book large language models, mobilize the willingness of users to use large language models, achieve high-quality user retention, and promote the non-human high-quality development of domain large language models. [Method/process]This study takes users in the field of ancient books as an example, based on resilience theory, and draws on grounded theory research methods to code and deconstruct first-hand data obtained from in-depth interviews with 30 users of ancient book large language models. The focus is on extracting influencing factors and conceptual categories from the perspective of resilience theory, and then constructing a research model on the mechanism of dropout behavior formation among users of ancient book large language models from the perspective of resilience theory. [Result/conclusion]The research results show that resilience factors, psychological resilience, cognitive factors, and situational factors are important factors affecting the dropout behavior of users of the ancient book large language model.Resilience factors include three dimensions:information resilience, technological resilience, and environmental resilience. Psychological resilience mainly goes through three stages: emotional stress, emotional resilience, and cognitive resilience.The research results provide necessary reference for effectively preventing users of ancient language models from dropout and achieving high-quality user retention.
  • WANG Dakun, HUA Bolin
    Scientific Information Research. 2023, 5(1): 92. https://doi.org/10.19809/j.cnki.kjqbyj.2023.01.007
    [Purpose/significance]With the help of information technology, quantitative analysis on policy text is an emerging interdisciplinary research direction.[Method/process]This paper systematically sorts out the current progress of quantitative research on policy text from the three-dimensional perspective of data sources, methods and applications. After summarizing the distribution of metadata and data sources of policy text, at the method level,  it is divided into three categories: content analysis method, bibliometric method and text mining method, and in the application of policy text mining, there are mainly policy topic mining, policy target tool mining, political position analysis, distribution of publishing agencies and policy diffusion research.[Result/conclusion]In the future, researchers should pay more attention to the mining of policy content and combine it with quantitative analysis research.
  • DUAN Yuzhu, FU Qiang, LI Yuqiong
    Scientific Information Research. 2025, 7(1): 86-94. https://doi.org/10.19809/j.cnki.kjqbyj.2025.01.008
    [Purpose/significance]The innovation of national defense-related science and technology promoting the development of new-quality fighting capacity has become a national strategic demand, effectively forecasting innovation pathways in national defense-related science and technology is of great significance to improve the innovation capability of national defense-related science and technology and create the Chinese new quality fighting capacity. [Method/process] In view of characteristics of the innovation of national defense-related science and technology, i.e. mutability, uncertainty and high risk, this thesis constructs a logical framework of forecasting innovation pathway in national defense-related science and technology based on the theoretical analysis of the core elements of forecasting innovation path, and adopts three submodels i.e. identification of military needs, prediction of key technologies and mining of supportive policies to forecast innovation pathway in national defense-related science and technology. [Result/conclusion]The research results show that the foundation model can effectively predict the development trend of national defense-related science and technology and key technological break throughs, which provide scientific basis for the decision-making of innovation of national defense-related science and technology and effectively promote the integration of the new quality productivity and the new quality fighting capacity.
  • XIE Danlin, HU Xisheng, YANG Weishu
    Scientific Information Research. 2024, 6(1): 90.
    [Purpose/significance]In the information age with decentralized discourse power, the channels for the public to publicly express their opinions and participate in topic discussions are increasing. The exchange and dissemination of online public opinion on low-ignition events will undoubtedly increase the heat of the event and bring greater pressure and challenges to corporate crisis public relations. [Method/process]Taking "Haitian soy sauce event" as an example, this paper discusses the stage of network public opinion dissemination of the event, uses ROST CM to carry out high-frequency word statistics and public sentiment tendency analysis, and uses Ucinet and Gephi to analyze the social network of the event from four dimensions including network density, network centrality, cohesion in group and point center. [Result/conclusion]This study found that the event public opinion has the characteristics of high public attention, large amount of discussion, large proportion of negative public opinion, and large influence of some opinion leaders. Based on the above characteristics, the countermeasures and suggestions of enterprise crisis public relations are put forward, in order to provide reference for enterprises to deal with the spread of crisis public opinion and eliminate negative effects.
  • GAO Jin-hu
    Scientific Information Research. 2020, 2(1): 12.
    [Purpose/significance]The thesis reviews the evolution of the intelligence analytic methodology and illuminates its development path,which aims at providing reference for the improvement in intelligence analysis.[Method/process]Through the elaboration of methodological evolution,the thesis proceeds to distinguish the national security intelligence analysis from the information analysis,summing up the influence of the analytic methodology on analysis.[Result/conclusion]Intelligence analysis is far more complex than conventional information analysis,while the latter is unable to cope with the intricate issues in national security intellgience analysis.Therefore,new analytic methodology should be constructed on the basis of falsificationism and cognitive psychology.
  • PENG Qining, LIU Bingxiang, FU Zhenkang, FENG Guangyu, BEI Wenyu
    Scientific Information Research. 2024, 6(1): 75.
    [Purpose/significance]Taking patent infringement as the starting point, this paper explores the influence mechanism of different factors on the tendency of patent infringement declaration, then compares and analyzes the differences in the influencing factors of invalid declaration under different infringement themes in the same field.[Method/process]Firstly, this paper uses the LDA topic model to subdivide the infringement topics in the selected emerging industry field, and understands the different infringement topics and infringement keywords of the infringement patents in this field;Secondly, the statistical correlation model is used to calculate various data indicators under different infringement classification topics and comparatively analyze the correlation between invalid declaration tendencies. Finally,by constructing a multi-feature fusion random forest model, the patents under different infringement classification topics are identified and trained for invalid declaration classification, and the LIME model in machine learning can be explained. Explain the degree of influence of the measurement index features in the model.[Result/conclusion]According to the correlation analysis after topic classification, it is found that under different topic classifications, the selected feature indicators not only have different overall influence on the determination of invalidity after infringement, but also have different influencing factors and influence degree rankings in different classification results. There existed significant differences in the classification rules and classification indicators relied on by different classification topics.
  • LAI Maosheng
    Scientific Information Research. 2024, 6(2): 1.
    [Purpose/significance]This paper attempts to sort out the development context of the field of knowledge organization, its evolution law and the driving factors of evolution, expounds the new needs and tasks of knowledge organization of scientific and technological information institutions in the era of big data and the innovative exploration of scientific and technological information community.[Method/process]This paper preliminarily investigates the research progress, challenges and shortcomings in the digitalization-intelligentize of knowledge organization in recent years, predicts the future development trend and ideas, and emphasizes the need to take the road of combining professional elite with grassroots, human intelligence and data intelligence. [Result/conclusion]Several areas that need to be paid attention to and strengthened in the development of knowledge organization in the library and information community are proposed, namely, data labeling, organization and retrieval of datasets, and metadata in data management.
  • WANG Yanfei
    Scientific Information Research. 2024, 6(3): 1-9. https://doi.org/10.19809/j.cnki.kjqbyj.2024.03.001
    [Purpose/significance]Integrity and innovation are important issues that the Institute of Science and Technology Intelligence needs to grasp. [Method/process]Regarding innovative issues based on adhering to principles and starting from historical facts, this article analyzes the role and significance of WIKID memes in scientific and technological intelligence study, explores relevant professional operational standards, proposes to strengthen the ability to compile WIKID memes under strategic concerns, implements intelligence thematic research based on dynamic clue discovery, and carefully examines the norms of meme characterization in perception scanning. [Result/conclusion]These are necessary actions for scientific and technological intelligence professionals and professional institutions to respond to the challenges of the times.
  • WEI Ruibin, WANG Yidan, XU Yan
    Scientific Information Research. 2025, 7(1): 41-52. https://doi.org/10.19809/j.cnki.kjqbyj.2025.01.004
    [Purpose/significance]The main path analysis of citation networks can be used to identify important literature in specific fields and can achieve the extraction of mainstream research threads. This paper will use the main path analysis method to analyze the research path of knowledge graphs and sort out the context of their research development. [Method/process]This paper firstly obtains research papers in the field of knowledge graphs from the Web of Science platform, then uses the HistCite software to generate a direct citation network of the literature, and then imports the data into Pajek to generate multiple main paths of the dataset, and combines the content of the papers on the main paths for qualitative analysis. [Result/conclusion]Through main path analysis, some main paths in the field of knowledge graphs can be quickly identified, such as the construction of knowledge graph, research on the application of knowledge graphs in recommendation and question answering and other application scenarios, research on the application of knowledge graphs in specific application fields such as manufacturing. These paths reflect the development context and research direction of knowledge graph technology. Review studies have played an important role in the development of the knowledge.
  • HUA Bolin, WANG Yingze
    Scientific Information Research. 2025, 7(1): 53-64. https://doi.org/10.19809/j.cnki.kjqbyj.2025.01.005
    [Purpose/significance]With the strong ability to process large-scale datasets and outstanding performance in various natural language processing tasks, large language models (LLMs) have excelled across multiple industries. Since scientific and technical intelligence primarily relies on textual data, LLMs are naturally well-suited for this field, ushering in a new wave of transformative changes. [Method/process]This article discusses the advantages of LLMs from five perspectives: low-dimensional dense vector representations of text, large-scale pre-trained models, fine-tuning and prompt learning, high-quality large-scale training data, and human alignment techniques. [Result/conclusion]LLMs have extensive applications in tasks such as intelligence identification, intelligence tracking, intelligence evaluation, and intelligence prediction, resulting in significant optimization improvements or paradigm shifts.
  • ZHU Danhao, ZHAO Zhixiao, WU Na, WANG Xiyu, SUN Guangyao, WANG Dongbo
    Scientific Information Research. 2024, 6(2): 11.
    [Purpose/significance]In this paper, we take the automatic text segmentation of ancient books as an entry point, introduce the "Xunzi" series of large language models, and explore the performance of large language models on the task of word division of ancient texts. [Method/process]This paper constructs an instruction dataset based on the Zuozhuan, with data cleaning and organisation.on this basis, 1 000 pieces were extracted from it as test data, then 500, 1 000, 2 000, and 5 000 pieces of data were used as training data to fine-tune the instructions and test their performance, respectively. [Result/conclusion]The experimental results show that only a relatively small amount of data is needed for the large language model to have a more desirable performance, and the Xunzi-Qwen-7B model shows optimal performance with an F1 value of 84.54% when the amount of fine-tuned data reaches 5 000 pieces.
  • SU Xinning
    Scientific Information Research. 2024, 6(1): 1-9. https://doi.org/10.19809/j.cnki.kjqbyj.2024.01.001
    [Purpose/significance]By analyzing the system and rules of traditional knowledge organization methods, the intelligent capabilities of traditional knowledge organization methods are refined and integrated into artificial intelligence(AI) technology, to enhance the precision and efficiency of AI in information processing. [Method/process]This paper reviews the development of knowledge organization and analyses the inherit structure and mechanisms of traditional knowledge organization methods. [Result/conclusion]Research suggests that over centuries of development and evolution, knowledge organization has gained the ability to reflect knowledge systems and disciplinary systems across different disciplines from diverse perspectives, establish semantic relations from diverse knowledge associations, and associate and integrate knowledge of different forms, types, and structures using scientific knowledge organization methods. These capabilities provide more effective ways for artificial intelligence to grasp knowledge systems, explore knowledge relations, reason about association probabilities in scientific problems, expand or refine knowledge and terms, as well as analyze relationships between things. The development of artificial intelligence in the field of information processing should be closely coordinated with knowledge organizations, to fully unleash its potential intelligent capabilities.
  • ZHAO Xueqin, YAO Yuhang, LI Tian'e, DONG Meiwen
    Scientific Information Research. 2023, 5(3): 26-35. https://doi.org/10.19809/j.cnki.kjqbyj.2023.03.003
    [Purpose/significance]With the development of the times, the concept of uncertain environment has also been paid more and more attention, and the multi-source heterogeneous uncertain environment has become a difficult problem hindering the integration and sharing of intangible cultural heritage digital resources.[Method/process]The integration of intangible cultural heritage digital resources knowledge elements based on evidence theory is proposed, and the integration framework of intangible cultural heritage digital resources knowledge elements is constructed, and empirical analysis is carried out based on the intangible cultural heritage digital resources in the Wanli tea ceremony.[Result/conclusion]The results show that the integration of knowledge attributes based on evidence theory can effectively utilize the value of intangible cultural heritage digital resources, and promote the digital preservation, development and utilization of intangible cultural heritage digital resources.