Intelligence Theories and Methods

255 The Knowledge Ecosystem for Key Core Technologies: Integration, Evolution, Synergy, and Impact Hot! Chu Jiewang, An Yiran, Li Jiaxuan DOI: 10.3772/j.issn.1000-0135.2023.03.001 The construction of the knowledge ecosystem is conducive to knowledge sharing, technology sharing, and talent exchange and enhances the capability of the technological innovation system. By exploring the synergy, evolution, integration, and impact of the knowledge ecosystem, the study can resolve the situation in which knowledge subjects are separated from each other and work in isolation, transform the traditional organization and implementation mode of tackling key problems, and collaborate with the key core technology tackling. This study combines the knowledge ecosystem and multiple helix theory, constructs the technology evolution path for key core technologies, the collaboration mode of knowledge subjects, analyzes the environmental impact on knowledge subjects, and finally forms the knowledge ecosystem according to the characteristics of key core technologies and the current situation of tapping. Finally, the study uses the Distribution, Interaction, Competition, and Evolution (DICE) model to explain the operation mechanism of the knowledge ecosystem, takes quantum information technology as an example to study its applicability, and provides a new idea for the construction of a new organizational model for key core technology tackling. 2023 Vol. 42 (3): 255-267 [Abstract] ( 332 ) HTML (181 KB)  PDF (2642 KB)  ( 195 )

268 Examining the Situation Awareness Model for Science and Technology under Intelligence Thinking Hot! Wang Wei, Yang Jianlin, Liang Jiwen DOI: 10.3772/j.issn.1000-0135.2023.03.002 Global competition in science and technology has increased the demand for effectiveness and currency in intelligence support. Establishing the situation awareness model could be an effective way for intelligence work to provide services for the development and security of science and technology. In this study, we first clarified the interrelationship and applicable scenarios of target-driven, data-driven, and task-driven intelligence. The process of intelligence awareness for science and technology was proposed by integrating the three driving forces. Subsequently, intelligence thinking was introduced for the demand of situation awareness in science and technology. We further clarified the requirement of system thinking, competitive thinking, all-sources thinking, and critical thinking in the intelligence process. Thereafter, the framework of intelligence awareness for science and technology—including the layer of situation awareness, situation understanding, situation prediction, and intelligence delivery—was constructed. The model proposed in this study could guide the realization of intelligence awareness and enhance the ability to support decision-making. 2023 Vol. 42 (3): 268-278 [Abstract] ( 341 ) HTML (97 KB)  PDF (3400 KB)  ( 268 )

279 Representative Paper Selection Based on Citation Comment Weighing Hot! Ma Ruimin, Liu Zhifang, Lyu Yuhan, Feng Yumei DOI: 10.3772/j.issn.1000-0135.2023.03.003 The scientific selection of representative paper is the basic work that ensures the implementation of the representative work evaluation system. It is theoretically and methodologically crucial to establish a novel alternative under the current background of “breaking the five-only.” This study first probes into the connotation of representative paper and the selection perspectives. Then it elaborates on the idea sources of selection models from theoretical and empirical perspectives. Furthermore, the selection model combing authority citer and citation comment classification is developed. Subsequently, the empirical research based on candidate projects of the National Natural Science Award in the field of physics is conducted. The obtained results demonstrate that the proposed selection model is a typical model of “selecting the best from best,” which can accurately select representative papers from several papers of applicants. Meanwhile, compared with non-self-citation count and impact factor, it further proves that the mode is superior — more discriminative and more stable. 2023 Vol. 42 (3): 279-288 [Abstract] ( 294 ) HTML (129 KB)  PDF (1760 KB)  ( 263 )

Intelligence Technology and Application

289 Knowledge Association and Aggregation of Large-scale Policy Texts Based on All-factor Network Construction Hot! Zhang Weichong, Wang Fang, Zhao Hong DOI: 10.3772/j.issn.1000-0135.2023.03.004 The current policy intelligent processing technology lacks deep correlation and effective aggregation of policy text corpus. This technical bottleneck makes it difficult to reuse policy knowledge and convert fragmented texts into systematic knowledge. Guided by theories of policy network, total factor network, and knowledge aggregation, and considering knowledge graph as its technical framework, this study proposes a technical solution for policy text association and aggregation based on total factor network. Using more than 160,000 policies as data samples, we realize the semantic representation and identification of policy texts, knowledge association and aggregation, policy network construction, and policy knowledge aggregation. The high correlation between policies is revealed from multiple dimensions, such as basis and theme. Moreover, large-scale application tests carried out using SARS and COVID-19 cases can provide reference for promoting policy science to effectively solve complex social problems. Finally, a platform prototype for policy intelligent analysis with information integration and network multi-dimensional analysis functions is designed to improve the practical applicability of the technical solution proposed in this study. 2023 Vol. 42 (3): 289-303 [Abstract] ( 410 ) HTML (131 KB)  PDF (7204 KB)  ( 157 )

304 Multi-dimension Public Opinion Mining of Social Media Based on the Hierarchical Viewpoint Tree Hot! Xi Haixu, Zhang Chengzhi, Zhao Yi, Tian Liang DOI: 10.3772/j.issn.1000-0135.2023.03.005 Mining of public viewpoints on social media can help people quickly and effectively understand them, avoiding subjective and casual comments and spreading wrong information, leading to malignant events. Currently, viewpoint mining on social media mainly analyzes public opinion from a single dimension such as the theme, tendency, or aspect content of viewpoints. It is difficult for people to fully understand public opinion and grasp multi-dimensional information such as the logical relationship between these viewpoints; therefore, the relevant performance of each subtask needs to be improved. To more accurately understand and comprehensively analyze public opinion information of different dimensions and promote people's in-depth understanding of public opinion on social media, this article proposes a construction method of the hierarchical viewpoint tree on social media which reflects the logical relationship between viewpoints in various dimensions, and selects the topic of hydroxychloroquine as the treatment of COVID-19 on Twitter to conduct an empirical study on this topic. The results show that the construction method of the proposed hierarchical viewpoint tree can provide multi-dimensional and understandable viewpoints on social media. 2023 Vol. 42 (3): 304-315 [Abstract] ( 310 ) HTML (140 KB)  PDF (2465 KB)  ( 306 )

316 Research on Event Extraction from Ancient Books Based on Machine Reading Comprehension Hot! Yu Xuehan, He Lin, Wang Xianqi DOI: 10.3772/j.issn.1000-0135.2023.03.006 Exploring the context of ancient Chinese classics and extracting the events and event arguments contained in ancient Chinese classics are critical to read and understand the content of the text quickly. At present, research on event extractions from ancient books is mainly based on pattern matching, machine learning, and neural networks. This paper integrates the machine reading understanding mode into the existing neural network-based methods and combines the “event type” and “argument role” in event extraction into the form of questions so that the answer is event argument. Zuo Zhuan (in annalistic style) and The Historical Records (in annal-biography style) are selected as the training and generalization data, respectively, and the confused sentences are introduced in the specific generalization process to verify the effect of the model, which provides a reference idea for ancient Chinese event extraction. 2023 Vol. 42 (3): 316-326 [Abstract] ( 350 ) HTML (117 KB)  PDF (1966 KB)  ( 363 )

327 Research on Author Name Collaborative Disambiguation Based on Meta-path Hot! Yang Zhao DOI: 10.3772/j.issn.1000-0135.2023.03.007 In the era of big data, facing the data governance of literature networks and accurately awarding academic achievements requires approaching the dual challenges of the diversity and ambiguity in author names. This paper proposes a meta-path method of author name disambiguation from a heterogeneous co-occurrence network perspective to solve the collaborative disambiguation problem, which arises from the coexistence of aliases, renaming, and the synonymous translations of author and institution names. The author name disambiguation problem is transformed into a heterogeneous network mining problem using the collaborative strategy of author and institution names. The author name disambiguation framework is constructed based on the meta-path. The heterogeneous author co-occurrence network model was constructed by combining the semantic and spatial association between objects. The variation aggregation, name disambiguation, and institution normalization methods based on the meta-path were proposed to calculate the similarity and perplexity of author and institution names. The effectiveness of this method was experimentally verified using the data set of English papers under the same institution and the data set of Chinese papers under the same name as examples. 2023 Vol. 42 (3): 327-340 [Abstract] ( 245 ) HTML (253 KB)  PDF (1710 KB)  ( 293 )

341 Analysis of Factors in Citing Scientific Papers in Policies against COVID-19 Pandemic Hot! Ren Chao, Yang Menghui, Li Kai, Yang Guancan, Lu Xiaobin DOI: 10.3772/j.issn.1000-0135.2023.03.008 In the context of the COVID-19 pandemic, the role of science in supporting policy decision-making has become increasingly important. Scientific papers have gradually become the main reference sources for policy making, and evidence-based decision-making guarantees the scientific development of policies against the COVID-19 pandemic. To explore the factors affecting the citation of scientific papers in policies and evaluate the importance of these factors from the perspective of evidence-based policy making, this study considers scientific papers as research objects and constructs an index system with three levels of influencing factors: primary indices (comprising research evidence factors, internal factors of researchers, and external environmental factors); secondary indices, comprising four important factors in citing scientific papers (namely rationality, originality, and scientific and social values); and 34 characteristics as tertiary indices. By combining five sampling and classification techniques, the accuracy of the influencing factors index system and prediction model reported here was verified. Additionally, the importance of the different factors was explored through feature importance and correlation analyses. The results show that the proposed factors index system, mainly at the three levels of papers, authors, and journals, are important factors affecting the citation of scientific papers in the policies against COVID-19. The results further show that the most important factors affecting the citation of scientific papers in policies are the academic influences of scientific papers and authors and the degree of dissemination of the papers in the news and social media. Concurrently, it was found that the mechanism of citing scientific papers in policy documents is relatively complex, and there is a need for further studies on multiple and complex factors influencing the selection of scientific papers as evidence-based reference sources in policy documents. 2023 Vol. 42 (3): 341-353 [Abstract] ( 305 ) HTML (146 KB)  PDF (2787 KB)  ( 238 )

354 Information Dissemination Model Used in Online Social Networks Based on Hypergraphs Hot! Shen Wang, Shi Qianru, Wang Junyao, Li He, Liang Shihao DOI: 10.3772/j.issn.1000-0135.2023.03.009 This study describes a network topology with hypergraphs based on pairwise friendship and virtual community friendship in online social networks. To improve the susceptible-infected-removal (SIR) information dissemination model, the individual states are defined after analyzing users’ forwarding and commenting behaviors, and the individual state transition rules are defined based on the social impact theory. The hypergraph network topology and the improved SIR model constitute the proposed network information dissemination model. The NetLogo software is adopted to simulate the proposed model and analyze the influence of the model parameters, such as the community structure. The simulation results show that the network structure based on hypergraphs fits well with the network structure of the virtual community. Moreover, the community structure, social network activeness, and information obsolescence speed have significant effects on the information dissemination process. 2023 Vol. 42 (3): 354-364 [Abstract] ( 485 ) HTML (192 KB)  PDF (2108 KB)  ( 299 )

Intelligence Users and Behavior

365 Construction of Information Problem Space to Foster Users' Information Need Expression Hot! Zhang Xiaoyue, Liu Chang DOI: 10.3772/j.issn.1000-0135.2023.03.010 Information seeking can be considered as a problem-solving activity. Understanding the construction of the users’ information problem space during information seeking is of great significance. Analyzing the complex, fuzzy, and dynamic changing states of users’ information needs from the perspective of a cognitive mechanism is beneficial for users to better solve information problems. Moreover, these solutions help in optimizing information systems and services. After systematically reviewing studies on information problem-solving theory and information need expression, the concept of an “information problem space” is proposed. This type of approach represents the topically faceted internal structure in a “purpose and means” manner. Then, based on the results of a simulated search experiment, the group user information problem space is illustrated. Suggestions for optimizing information systems are discussed. The results show that the proposed information problem space can represent the dynamic cognitive process of users’ information search process. Future research should focus on representing the user information problem space (including visualization approaches) and the interactive support function of search systems for users to construct information problem spaces. 2023 Vol. 42 (3): 365-379 [Abstract] ( 230 ) HTML (158 KB)  PDF (2364 KB)  ( 357 )