Intelligence Theories and Methods

1129 Exploring and Anticipating Early Detection Methods for Scientific and Technological Weak Signals Hot! Zhang Huiling, Xu Haiyun, Liu Chunjiang, Chen Liang, Wang Chao, Wang Haiyan DOI: 10.3772/j.issn.1000-0135.2024.10.001 Weak signals in science and technology (WSST) are indicators of future technological advancements, offering insights into innovation trends, emerging opportunities, and market directions. However, there is currently a lack of comprehensive method libraries for early WSST perception. Existing approaches are limited in their ability to mine weak technological signal features and suffer from low interpretability and subjective constraints. This article systematically categorizes potential models and methods for WSST perception, with the goal of enabling intelligent analysis for technology foresight. We begin by examining the essence of WSST and the attributes of intelligent perception. Then, we integrate three methodological dimensions—quantitative and qualitative, subjective and objective, and causal and correlative—to compile a method library and toolset for early WSST perception. By assessing WSST characteristics, applicability, and interpretability, we identify methods suitable for various stages of signal perception. Additionally, we evaluate intelligent perception techniques, highlighting their advantages and limitations, while proposing future research directions. Case studies in the field of stem cells validate the feasibility of the proposed perception methods and processes. Moving forward, there is a need for novel cognitive perception models and traceability methods for WSST, aimed at expanding characteristics coverage and developing perception techniques driven by digital intelligence to support emerging technologies in uncertain environments. 2024 Vol. 43 (10): 1129-1141 [Abstract] ( 96 ) HTML (150 KB)  PDF (3506 KB)  ( 124 )

1142 Construction of Key Liaison Scholar Identification Model from the International Scientific Research Cooperation Network Perspective Hot! Guo Yusen, Yang Yanping DOI: 10.3772/j.issn.1000-0135.2024.10.002 International scientific research cooperation is an essential driver of global technological advancement and innovation. Scholars who play liaison roles are crucial to the construction and maintenance of these networks. Identifying key collaborators and establishing significant cooperative relationships in an increasingly competitive global scientific collaboration environment is a key research focus in information science. This study aimed to develop a model for identifying key liaisons using social network analysis methods. After segmenting the scientific research cooperation network into communities using the Louvain algorithm, three indicators were introduced to identify different types of liaisons: intra-community participation index Z, inter-community participation index P, and community connectivity index C. These results provide theoretical support for policies of China on introducing academic talent and seeking partners in global scientific research cooperation. The research findings will not only help scientists discover important international partners and promote knowledge sharing and interdisciplinary communication, but also facilitate the development of joint international research projects and provide information for the introduction and cultivation of international talent. 2024 Vol. 43 (10): 1142-1153 [Abstract] ( 87 ) HTML (167 KB)  PDF (3763 KB)  ( 156 )

1154 Metamodel Study of Knowledge-Organization-System Theory Integrating Multidisciplinary Theories Hot! Zhai Kexin, Yuan Jingshu, Yuan Man DOI: 10.3772/j.issn.1000-0135.2024.10.003 Based on literature review, most current studies pertaining to knowledge organization systems (KOS) focus primarily on the construction of domain-knowledge topic models, whereas few studies begin from the underlying fundamental theories to reveal the basic principles and operation mechanisms of KOSs. This results in inadequate guidance pertaining to the underlying methodology of KOS research and application, thus rendering it difficult to guarantee the quality and shareability of KOSs. Hence, this study first proposes a three-layer framework model that addresses the grammatical, semantic, and pragmatic aspects of KOSs. Based on this model, the operational mechanisms of the conceptualization process and the symbolic representation of knowledge semantics in knowledge organization are revealed. Second, by focusing on the semantic layer while integrating multidisciplinary theories, such as semantic triangulation, concepts, semiotics, and system theories, a theoretical metamodel of the knowledge-organization system is proposed, in addition to the method and process used for constructing KOSs in the domain. Finally, the feasibility of the theoretical metamodel and the KOS construction method and process is verified by considering knowledge-model construction in petroleum underground operations as the research object, thus providing a new scientific method for the high-quality and high-efficiency construction of KOSs in other fields. 2024 Vol. 43 (10): 1154-1165 [Abstract] ( 75 ) HTML (105 KB)  PDF (2603 KB)  ( 88 )

1166 Quantitative Identification and Influencing Factors of Sleeping Beauty Papers in an Open Science Environment Hot! Wang Xu, Xue Yufei, Qiu Junping DOI: 10.3772/j.issn.1000-0135.2024.10.004 The rapid development of social media has profoundly influenced and transformed the modes of scientific communication. The rise of open science as a new research paradigm provides a new perspective and approach for studying the characteristics and influences of sleeping beauty papers. The quantitative identification and analysis of influencing factors based on multisource heterogeneous data can fully tap into the potential research value of open science. This study considers economics and computer science as representative disciplines in the social and natural sciences, respectively. Using massive citation data and altmetrics data obtained from the Web of Science and Altmetrics.com, we identified sleeping beauty papers published in the two disciplines from 1996 to 2014. To examine and compare the influencing factors of sleeping beauty papers, descriptive analyses were conducted from three aspects: sleep characteristics, wake-up mechanisms, and journal characteristics. The following conclusions were drawn. (1) It is necessary and meaningful to combine the citation frequency of papers reflecting academic attention with altmetric data reflecting social attention as raw data for identifying sleeping beauty papers. (2) Natural sciences are more prone to the emergence of sleeping beauty papers with clear sleep characteristics. Although types of research focus on the theoretical innovations and values of articles, natural science places greater emphasis on the application and improvement of existing algorithms and technologies. Papers published in high-level journals in the social sciences field may also face the situation of being awakened after years of dormancy, whereas sleeping beauty papers in the natural sciences field often appear in low-level journals. (3) Journal influence, news coverage, and social media reading were discovered to have a significant impact on the awareness of foreign sleeping beauty papers in the field of social sciences. In contrast, funding had a significant impact on the emergence of sleeping beauty papers in the natural sciences. Policy texts play a key role in both fields. (4) Compared to the achievements of foreign resource library construction, China should continue to develop altmetric data integration platforms that are adapted to local conditions. 2024 Vol. 43 (10): 1166-1181 [Abstract] ( 69 ) HTML (177 KB)  PDF (5082 KB)  ( 116 )

1182 Mechanism of Knowledge Fusion and Value Transmission in Structural Unit of Citation-Network Knowledge Fusion Hot! Chen Yongyue, Xu Chunmeng, Xu Tingmei DOI: 10.3772/j.issn.1000-0135.2024.10.005 This study analyzes the process of knowledge fusion and the mechanism of value transmission in citation networks from a micro-perspective as well as reveals and evaluates the value, role, and contribution of literature in citation networks from a quantitative perspective. In this study, the basic structural units of knowledge-convergence fusion and knowledge-diffusion fusion are decomposed from a citation network. By representing and quantifying knowledge stock, knowledge flow, knowledge-fusion amount, and knowledge-fusion degree, an analysis method for the values of knowledge reference, knowledge flow, and knowledge fusion is constructed. Representative literature is selected from the Web of Science database for experimental analysis. Using quantitative data, the values of knowledge citation, knowledge flow, and knowledge integration are calculated, and the value, role, and contribution of the literature in knowledge integration are analyzed. Combining citation positions and the functional structure of papers, this study identifies the main literature that provides innovation and citations, as well as highly compatible and strongly correlated cited and citing literature. 2024 Vol. 43 (10): 1182-1198 [Abstract] ( 75 ) HTML (363 KB)  PDF (1471 KB)  ( 96 )

Intelligence Technology and Application

1199 Essential Reference Measurements from the Perspective of Full-Text: Concept Definition, Index System, and Identification Model Hot! Lin Gege, Hou Haiyan, Pan Yuxin, Liang Guoqiang, Hu Zhigang DOI: 10.3772/j.issn.1000-0135.2024.10.006 Identifying essential references within citing documents is fundamental for conducting thorough evaluations of scientific achievements. Therefore, this study explores the measurement of essential references from the perspective of full text that includes the definition of concepts, construction of an indicator system, and optimization of identification models, thereby providing a more precise scientific evaluation tool. First, the definition of essential references was clarified, and an indicator system for identifying essential references was constructed, encompassing two dimensions (bibliographic and citation information), eight sub-dimensions, and 33 citation feature indicators. Second, by utilizing various machine learning models, such as random forest, support vector machine, and logistic regression, citation feature indicators were selected and optimized. Their correlations and information gains were analyzed, and 21 important citation feature indicators were retained, to validate the effectiveness of the identification models. The results indicate that citation feature indicators based on citation information hold greater importance and contribute more to the identification of essential references. The performance of machine learning models in identifying essential references was excellent, particularly for the random forest, support vector machine, and logistic regression models, with area under receiver operating characteristic curve (AUC) values exceeding 0.85, demonstrating the efficiency and robustness of the models. The core citation measurement methods and identification models not only provide more accurate tools for scientific evaluation systems but also lay a solid foundation for further in-depth research into citation analysis. 2024 Vol. 43 (10): 1199-1212 [Abstract] ( 89 ) HTML (137 KB)  PDF (4269 KB)  ( 94 )

1213 User Identification Across Social Media Based on Heterogeneous Graph Attention Network and Multi-Feature Fusion Hot! Bi Datian, Zhang Xue, Kong Jingyuan, Chen Gongkun DOI: 10.3772/j.issn.1000-0135.2024.10.007 Cross-social media user identification is crucial for guiding the collaborative governance of online public opinion and for comprehensively identifying and predicting user preferences. This study introduces a model for recognizing cross-social media users that integrates heterogeneous feature embedding and dynamic entity association to address the issues of weak data representability and neglect of the dynamic and associative nature of user information in current methods. The heterogeneous information networks of various social media platforms were created by incorporating basic user attributes, content generation, and social structure information. A new meta-path recognition strategy was designed to construct an adjacency matrix, allowing the heterogeneous graph attention network model to aggregate user node information and enhance the representability of node features. Additionally, three continuous time decay functions were introduced to weigh the entity similarity matrix across social media platforms, enhancing the dynamic relationship between entities. By integrating features from both single and cross-social networks, a multi-layer perceptron was utilized to achieve the recognition and prediction of cross-social media users. Experiments conducted on the real Weibo-Zhihu dataset showed that the overall performance of the model was superior to that of other benchmark models. The linear attenuation function was found to have the most significant impact, and the meta-path detection strategy proposed in this article played a pivotal role in improving detection effectiveness. 2024 Vol. 43 (10): 1213-1226 [Abstract] ( 90 ) HTML (211 KB)  PDF (2919 KB)  ( 139 )

1227 Topic Transition Paths and Risk Assessment of Online Public Opinion in Public Emergencies Hot! Zhou Wei, An Lu, Han Ruilian DOI: 10.3772/j.issn.1000-0135.2024.10.008 Exploring the topic transition nodes of online public opinion in public emergencies and conducting multi-phase risk assessments are of great significance for accurately addressing crises of online public opinion in these cases and providing dynamic guidance strategies. This paper proposes a method for identifying topic transition paths and conducting a multiphase risk assessment of online public opinion in emergency situations. First, a temporal semantic co-occurrence network was constructed by integrating the RoBERTa model, and network community topics were discovered using the Louvain-CFDP algorithm. Second, a model for detecting topic transitions was developed to generate topic transition paths and identify and analyze multiple types of transition paths and risk fluctuation characteristics. Taking the “Japanese nuclear contaminated water discharge into the sea” incident as an example, the empirical analysis identified three types of transition paths: event development, emotional aggregation, and derivative event types. The characteristics, risk features, and differences between the three types of paths were analyzed. The results show that the proposed method for topic transition paths and risk assessment can comprehensively demonstrate the topic transition of emergencies on social media, provide guidance and references for government departments to quickly identify high-risk topics, and formulate precise and effective public opinion risk guidance schemes. 2024 Vol. 43 (10): 1227-1241 [Abstract] ( 90 ) HTML (193 KB)  PDF (5858 KB)  ( 157 )

Intelligence Users and Behavior

1242 Health Information Use of Chronic Disease Patients: An Analysis of the Influence of Individual Characteristics and Health Information Characteristics Hot! Li Yuelin, Fan Sinuo, Zhang Yao DOI: 10.3772/j.issn.1000-0135.2024.10.009 Based on prior qualitative research, this study aimed to reveal how patients with chronic diseases use health information in their health management, and the factors affecting the usage of health information at different levels (high or low). A questionnaire survey was conducted to collect data from patients with chronic diseases. Multiple linear regression analysis and fsQCA were used to analyze the data. The findings indicate that the relevant factors influence the use of health information through two patterns: individual-characteristics-driven and information-features-driven. The individual-characteristics-driven pattern includes “cognitive-ability-driven,” “motivation-driven,” and “dual-factor-driven” patterns, with users' health information literacy and self-efficacy as the core influencing factor. The information-features-driven pattern covers “content-quality-driven” and “multi-factor-driven” patterns. The core factors for the “content-quality-driven” pattern are argument quality and emotion tendency. The core factors for the “multi-factor-driven” pattern are source professionality, credibility, actionability, and visual information medium. This study expands the field of health information behavior studies by developing a theoretical model of chronic patients’ health information use based on influencing factors. This study provides theoretical and practical guidance for improving health information services targeted at patients with chronic diseases. 2024 Vol. 43 (10): 1242-1256 [Abstract] ( 71 ) HTML (229 KB)  PDF (1401 KB)  ( 135 )