Intelligence Theories and Methods

381 Pattern Mining Based on Attribution Analysis and Its Empirical Study Hot! Cui Yunxue, Wang Xianwen, Wang Yongzhen DOI: 10.3772/j.issn.1000-0135.2023.04.001 The citation patterns of academic literature exhibit various citation motives, which restrict the in-depth understanding of researchers’ citation behavior. To address this issue, this study took attribution analysis as its research perspective and collected 500,000 citation relationships from PubMed Central as the research sample to reveal the composition of reasons behind citation patterns and quantitatively explain them. First, 12 citation reasons were selected as features to characterize the citation relationships from the academic and non-academic motives of citation. Second, a decision forest algorithm was used to conduct classification experiments on 500,000 real citation relationships and an equal number of paired virtual citation relationships based on the constructed features. Finally, the experimental results were attributed through the SHapley Additive exPlanations (SHAP) explanatory framework, evaluating the influence and mode of action in the 12 citation reasons in citation decisions. The empirical results indicate that the roles played by different citation reasons in deciding whether to cite an article vary considerably. Specifically, factors such as the relevance of a topic, similarity of the research context, and academic level of the cited author play a major role in the citation decision, while other factors such as journal influence and topic frontiers play a minor role. In addition, different citation reasons function in different ways on citation decisions, and the relationship between the change in the value of their features and degree of influence on citation decisions can be summarized into four types: S-curve, logarithmic growth, dichotomous, and random fluctuation. 2023 Vol. 42 (4): 381-392 [Abstract] ( 254 ) HTML (118 KB)  PDF (2196 KB)  ( 154 )

393 Demystifying Knowledge Growth of Interdisciplinary Fields Based on the Location Functions of Terms Hot! Cao Yujie, Wang Shiyun, Mao Jin, Li Gang DOI: 10.3772/j.issn.1000-0135.2023.04.002 The law of knowledge creation and flow in the formation and development of interdisciplinary fields can be explored by mining the interdisciplinary literature of scientific knowledge. Based on the location functions of terms in the literature, we propose literature spatial theory and a knowledge growth process model for interdisciplinary fields, which involves knowledge absorption, internalization, and innovation. We designed a full-text analysis method for quantitatively analyzing the knowledge growth process of interdisciplinary fields. Empirical research was conducted on the bioinformatics field, and the knowledge growth characteristics of this interdisciplinary field were determined. First, knowledge internalization and absorption are highly related; they vary in quantity but have the same trend. Second, the first peak of knowledge innovation appears four years later than knowledge absorption and internalization. Third, with the development of the interdisciplinary field, the real-time internalization rate remains relatively stable, the entire internalization rate decreases, and the internalization rate of newly absorbed knowledge is decreasing while the efficiency of knowledge innovation stimulated by knowledge internalization is increasing. The proposed full-text analysis framework for interdisciplinary knowledge growth can enrich the full-text content analysis methodology of academic literature. 2023 Vol. 42 (4): 393-406 [Abstract] ( 258 ) HTML (143 KB)  PDF (4987 KB)  ( 230 )

407 Interdisciplinary Research Modes and Paths Mining in Scientific Research: From the Scholars' Perspective of Changing NSFC Code Hot! Wu Xiaolan, Guo Yujie, Jiang Ning DOI: 10.3772/j.issn.1000-0135.2023.04.003 Owing to the requirements of scientific knowledge and social development, interdisciplinary research has become important in scientific research. During the NSFC project, the same scholar chooses different fund codes at different times, resulting in a co-occurrence relationship between different discipline application codes. This co-occurrence relationship proves, to some extent, the existence of interdisciplinary research behavior. Therefore, based on the code transformation perspective of NSFC, in this study, the interdisciplinary research modes and transfer paths have been studied. First, combining the hierarchical structure of NSFC discipline application codes, the diversity of individual research disciplines (DIRD) has optimized and then is applied to the interdisciplinary research mode and interdisciplinary occurrence path. Through research, it was found that the optimized measurement indicators can more accurately represent the diversity of individual research disciplines than the existing indicators. Moreover, a significant correlation exists between individual interdisciplinarity and whether the institutions belong to the “985 Project.” The number of scholars engaged in interdisciplinary research is also related to local economic level. In addition, there are typical scientific department transfer paths in project applications, and there exist obvious knowledge communities among 89 first-class disciplines. The results of this study not only further enrich the interdisciplinary measurement indicators of fund project data but also provide a useful reference for helping scholars to apply for interdisciplinary projects. 2023 Vol. 42 (4): 407-419 [Abstract] ( 132 ) HTML (174 KB)  PDF (2260 KB)  ( 232 )

Intelligence Technology and Application

420 Graph Neural Network-based and Particle Swarm Optimization Technological Prediction Model Hot! Lian Zhixuan, Wang Fang, Kang Jia, Yuan Chang DOI: 10.3772/j.issn.1000-0135.2023.04.004 Effective prediction of technological development trends is crucial for policymaking in many technological industries. As patent application forms express technical features in an in-depth manner, they can be used to train a model to improve the technology prediction. This study constructs a technology prediction model on the basis of particle swarm optimization (PSO) and graph neural networks to help improve the prediction accuracy of the future development trend of a technology field and the characteristics of its emerging technologies. Using 594 artificial intelligence patent applications arranged in China by US companies over the last two decades as the research objects, this study conducted an experiment and found that the suggested model obtains higher accuracy than baseline algorithms, such as the exponential smoothing method, moving average method, support vector regression, gate recurrent unit, and recurrent neural network. Moreover, the model can reveal the formation process of the new characteristics of the technologies. The study predicts the layout of US patents in China to investigate the trend, characteristics, and gaps in the current US technology layout in China. The proposed graph neural network-based and PSO technological prediction model can improve the technology prediction accuracy and support the decision making in the layout of technological industries and the funding of scientific research. 2023 Vol. 42 (4): 420-435 [Abstract] ( 308 ) HTML (203 KB)  PDF (7283 KB)  ( 246 )

436 Development of China's Public Data Governance Policy Based on Quantitative Textual Analysis Hot! Zhou Yi, Chen Bikun, Ma Jianghua, Sun Jinzhou, Shen Jiajun DOI: 10.3772/j.issn.1000-0135.2023.04.005 Public data are data collected and generated by public institutions in the process of performing their duties. Multichannel governance of public data, including how it is collected, processed, shared, and opened, is an important guarantee for harnessing the power of such data. Analyzing the characteristics, internal laws, and basic trends of development and change in data governance policies in China can provide insight for the improvement and innovation of these policies. Based on a comprehensive examination of Chinese data governance policy texts, this paper uses statistical analysis and text mining to extract and identify the characteristics of these texts. It focuses on high-frequency phrases, changes in policy contents, and keyword distributions of 258 data governance policies. The objective of data governance policies constantly change as the scope of its coverage expand. Furthermore, data governance policies are beginning to pay more attention to the rights of relevant subjects, data development and utilization, and the role of data in economic and social development. 2023 Vol. 42 (4): 436-452 [Abstract] ( 421 ) HTML (154 KB)  PDF (4534 KB)  ( 273 )

453 Technology Convergence Path and Effect of China's Solar Energy Industry Driven by Core Technology Hot! Liu Li, Su Lifang, Lou Xuming, Cheng Long DOI: 10.3772/j.issn.1000-0135.2023.04.006 The core technology provides technical opportunities for reversing the plight of enterprise innovation and achieving technological breakthroughs. Researching the technology convergence path driven by core technology can help innovation subject to clearly define the trajectory of technological development so that R&D resources can be allocated reasonably in different technological fields, which would improve innovation performance. We use the NPCIA core technology identification framework, taking the patent data of China’s solar energy industry from 2004 to 2018 as the research object, analyzing the core technology-driven technology convergence path based on the perspective of knowledge flow, and using LMDI model to test the driving effect of core technologies from four aspects: technology breadth, cross-integration strength, technology scale, and technology convergence depth. The results show that the core technology-driven technology convergence path includes the core path, edge absorption path, and edge diffusion path. Technology breadth and technology scale positively affect the technology convergence. Technology convergence depth and cross-integration strength exhibit a two-way trend. Among them, the driving effect of cross-integration strength is the strongest. 2023 Vol. 42 (4): 453-464 [Abstract] ( 214 ) HTML (168 KB)  PDF (3749 KB)  ( 136 )

Intelligence Discipline Development and Construction

465 Data Intelligence Empowerment and the Cultivation of Intelligence Thinking under the Background of Data Intelligence Empowerment Hot! Shen Shujing, Yang Jianlin DOI: 10.3772/j.issn.1000-0135.2023.04.007 This paper discusses the connotation of “Data Intelligence Empowerment” and the cultivation of intelligence thinking under the background of “Data Intelligence Empowerment.” It provides corresponding suggestions and references for intelligence studies to better adapt to the development of the times and contributes to the efficient realization of “Data Intelligence Empowerment.” This study analyzes the basic problems of “Data Intelligence Empowerment” from the perspective of intelligence studies, clarifies the key and difficult problems in the practice process of “Data Intelligence Empowerment,” and elucidates the new requirements of “Data Intelligence Empowerment” for the cultivation of intelligence thinking; in addition, it explores the practice path of intelligence thinking training under the background of “Data Intelligence Empowerment.” The study demonstrates that the process of “Data Intelligence Empowerment” requires the implementation of intelligence thinking to ensure the quality and efficiency of “Data Intelligence Empowerment.” Based on the orientation and original intention of intelligence talent training, it is necessary to clarify the practical value of intelligence thinking cultivation from the perspective of top-level logic, with the theoretical exploration of the new requirements of intelligence thinking for “Data Intelligence Empowerment” as the theoretical support, and logically integrate complete intelligence thinking into the teaching, application, and practice in the field of intelligence studies for the actual needs of “Data Intelligence Empowerment” and national strategies. 2023 Vol. 42 (4): 465-476 [Abstract] ( 339 ) HTML (106 KB)  PDF (1738 KB)  ( 381 )

Intelligence Reviews and Comments

477 Recent Advancements in Detection and Evolutionary Tracking of Scientific Topics: A Multi-perspective Survey and Prospect Hot! Cen Yonghua, Wang Yuefen DOI: 10.3772/j.issn.1000-0135.2023.04.008 Leveraging instructive approaches to comprehensively and accurately detect and track topics from the big data of historical literature is a hot and cutting-edge issue that has garnered substantial interest from different disciplines, especially scientometrics in information science. The major underlying mechanisms involve four perspectives, which include frequency, content, citation, and synthesized analyses. This study attempts to review the latest literature published in prestigious international and domestic journals, to update the main progress of relevant perspectives in scientific topic detection and evolution analysis by unravelling their essential implementation paths. In particular, it considers the risks inherited in existing perspectives, including the biases in heterogeneous importance of knowledge units or network relationships, temporal decay of knowledge, trap of the small samples of emergent topics, dilemma in fitting the natural development and evolution of topics, and failure in portraying knowledge flow and evolution in micro level. In response to these issues, the study focuses attention to the future trend of a synthesized perspective. 2023 Vol. 42 (4): 477-494 [Abstract] ( 263 ) HTML (247 KB)  PDF (2181 KB)  ( 395 )

495 AI Human-Computer Interaction of Intelligence Recommendation System: Frontier and Future Agenda Hot! Wang Xiwei, Wuji Siguleng, Liu Yutong, Luo Ran DOI: 10.3772/j.issn.1000-0135.2023.04.009 In terms of academic research and industrial application of artificial intelligence (AI), intelligent recommendation (IR) has become a key research issue. Analysis of the AI human-computer interaction research frontier and future agenda for IR can better promote interdisciplinary scholars to conduct further in-depth and extended research in this field, and understand the latest progress of information behavior research from the perspective of users. This study adopts the grounded literature review method to analyze the literature from China National Knowledge Infrastructure, Web of Science, and Association for Computing Machinery databases. A total of 64 documents were rigorously and effectively selected by defining and searching research questions, selecting document collections, selecting coding, and displaying results. The aforementioned research frontier and future agenda are comprehensively analyzed. The research frontier focuses on AI human-computer interaction behavior and influence for IR, perception and emotion expression, and scenes and service applications. In the future, scholars can conduct in-depth and interdisciplinary collaborative research around four aspects: new AI human-computer interaction relationship, form, influence, and equipment for IR. 2023 Vol. 42 (4): 495-509 [Abstract] ( 271 ) HTML (182 KB)  PDF (2556 KB)  ( 757 )