Special Topics

1 Concepts, Classification, Method Analysis, and Improvement Strategies of Science and Technology Evaluation Hot! He Defang, Pan Yuntao DOI: 10.3772/j.issn.1000-0135.2023.01.001 To address the complexities, uncertainties, and non-consensus of science and technology evaluation, this study, starting from the origin and purpose of evaluation, clarifies the concepts, classification, and functions of science and technology evaluation and proposes the use of corresponding reasonable evaluation methods. The study aims to construct a scientific and effective evaluation system and mechanism and establish an evaluation-oriented evaluation system with academic contribution and innovation value at its core. To create a healthy and sustainable innovation ecology, the study presents improvement strategies for science and technology evaluation in terms of special evaluation scheme formulation, coordination of different classifications of evaluation, full-process evaluation, evaluation of true problems, innovative methods, quantitative evaluation, scientific evaluation tool development, academic evaluation of science and technology personnel, and open evaluation. 2023 Vol. 42 (1): 1-9 [Abstract] ( 546 ) HTML (83 KB)  PDF (924 KB)  ( 319 )

Intelligence Theories and Methods

10 A Study on the Potential Collaborative Discovery of Industry-Academia-Research Based on Patent Documents Hot! Fang Siyue, Chen Fang, Wang Xuezhao DOI: 10.3772/j.issn.1000-0135.2023.01.002 Using patent documents to construct an industry-academia-research collaboration network and identify potential industry-university-research collaboration can help increase efficiency. A link prediction and coupling analysis method based on patentee manual codes were introduced into the collaboration network. The similarity index of link prediction was used to calculate the path similarity of the patentee and cosine distance was used to calculate the content similarity of the patentee. A weighted fusion index was constructed by fusing path similarity and content similarity. Area Under Curve (AUC) was used to determine the weight of the fusion index, and the effect of the index was tested in the industry-academia-research collaboration network for the biopharmaceutical industry, using data from the years 2014 to 2018. The empirical results showed that when the ratio of path similarity and content similarity was 1∶9, the prediction result performed the best. The potential collaboration results predicted by the optimal algorithm can be used to support the decision-making of industry-university-research innovation subjects on future cooperation partners in the biopharmaceutical industry. 2023 Vol. 42 (1): 10-18 [Abstract] ( 321 ) HTML (156 KB)  PDF (1313 KB)  ( 293 )

19 Scientific Breakthrough Topics Identification in an Early Stage Using Multiple Weak Linkage Fusion Hot! Liu Yahui, Xu Haiyun, Wu Huawei, Liu Chunjiang, Wang Haiyan DOI: 10.3772/j.issn.1000-0135.2023.01.003 By the term “scientific breakthroughs,” we mean innovations that are relatively transformative and have a profound impact on the future direction and trends of a disciplinary field. A scientific breakthrough aims to reveal new phenomena and laws that have emerged earlier but have not yet been recognized or investigated formally. Identifying scientific breakthroughs at an earlier stage can provide decision support to policymakers and granting agencies in optimizing resource allocation. This study selected the field of Gene Engineered Vaccine for empirical research—focusing on weak association linkages in knowledge networks—and constructed a multi-layer network based on subject term co-occurrence, author co-authorship, and reference co-citation relationships. Thereafter, we analyzed the association information between multiple feature items to mine the subject content and evaluated the recognition effect with the help of expert judgment and authoritative reports. The comparative analysis with the identification results based on strong correlation relationships verified that the method constructed in this study is applicable to the early identification of scientific breakthroughs. Future research could draw on network representation learning and introduce temporal networks to pinpoint the moment when a breakthrough is made to fade away. 2023 Vol. 42 (1): 19-30 [Abstract] ( 188 ) HTML (156 KB)  PDF (2292 KB)  ( 302 )

31 A Comparative Study of the Academic and Social Impact of Papers from the Diffusion Perspective: Taking Biomedicine as an Example Hot! Zhang Jingwen, Min Chao DOI: 10.3772/j.issn.1000-0135.2023.01.004 The dynamic process of citation diffusion demonstrates more comprehensively the quality and influence of scientific papers. From the perspective of citation diffusion, we introduce Altmetrics and the traditional citation indicators and attempt to comprehensively explore the impact of research under different citation diffusion patterns and the correlation between the indicators of academic and social impact. This study suggests that the Altmetrics score in regular and continuous citation diffusion patterns is positively correlated with citations, and these patterns have both high academic and social impact. In an irregular citation diffusion pattern, citations are not related to social media, indicating that high social impact does not always represent high academic impact. In the initial period of publication, the Altmetrics score can promote the citation. In this study, we analyze the research impact of different citation diffusion patterns with Altmetrics and citations from the perspective of diffusion, thereby providing a reference for evaluating the impact of scientific research using multi-dimensional indicators, as well as for improving academic impact. 2023 Vol. 42 (1): 31-42 [Abstract] ( 271 ) HTML (151 KB)  PDF (1443 KB)  ( 281 )

43 Effect of the Diversity of Scientific Teams on Disruptive Innovation in Academia: A Case Study in the Field of Artificial Intelligence Hot! Tang Xuli, Li Xin DOI: 10.3772/j.issn.1000-0135.2023.01.005 Although disruptive innovation in academia has become a new engine for technological progress and economic development, few researchers have studied the factors that affect disruptive innovation in academia. In this study, we take scientific teams as the research object and systematically analyze the effect of diversity in scientific teams on disruptive innovations in academia from three perspectives: social category, information, and behavior willingness. First, we collected collaboration papers on artificial intelligence (AI) published between 1950 and 2019 from the Microsoft Academic Graph. We extracted and represented author characteristics using topic model and text mining to compute the diversity indicators of scientific teams. Additionally, we measured disruptive innovation at the article level using the disruptive index. Then, the correlation analysis and ordinary least squares (OLS) regression analysis were used to eliminate the diversity indicators unrelated to disruptive innovation in academia. Finally, coarsened exact matching was used to explore the causal relationships between the diversity of scientific teams and disruptive innovation in academia. The results show that in the field of AI, the topic diversity and nation diversity of scientific teams have a significant causal relationship with disruptive innovation in academia. The degree of disruptive innovation in academia will decline 8.776%-19.000% when the topic diversity of scientific teams increases; it will also decline 5.493%-7.693% when the nation diversity increases. In addition, the results also show that there is no causal relationship between the diversity of behavior willingness of a scientific team and the degree of disruptive innovation in academia in the field of AI. 2023 Vol. 42 (1): 43-58 [Abstract] ( 296 ) HTML (346 KB)  PDF (1330 KB)  ( 393 )

Intelligence Technology and Application

59 Misinformation Identification Method by Automatic Iterative Clustering Data Set for Training Hot! Zhang Junsheng, Sun Xiaoping, Liu Zhihui DOI: 10.3772/j.issn.1000-0135.2023.01.006 With increasing proliferation of misinformation on the Internet, automatic identification of misinformation has become an urgent need for information governance. Misinformation on the Internet is constantly generated with new events, thereby resulting in the need for iterations and updates in the machine learning model to identify such misinformation. A new training data set should be constructed for each iteration update, so that the new misinformation can be reflected in the training set. Therefore, this study proposes a misinformation recognition method of dynamically and iteratively updating the training set to build a machine learning model, and iteratively clustering the misinformation data set based on kernel density estimation. In each cluster, training set and test set samples are selected to construct the corresponding classifier training data set and test data set; this enables the samples of new events to be reflected in the training set. The experimental results show that the misinformation classifier trained by the iterative clustering method based on kernel density estimation can significantly improve the accuracy of false information classification compared with the random data set division strategy. 2023 Vol. 42 (1): 59-73 [Abstract] ( 271 ) HTML (190 KB)  PDF (6073 KB)  ( 226 )

74 Smap: Visualization of Scientific Knowledge Landscape Based on Document Semantics Hot! Zhang Shuang, Liu Feifan, Luo Shuangling, Xia Haoxiang DOI: 10.3772/j.issn.1000-0135.2023.01.007 Given the explosive growth of academic literature, the continuous cross-fusion of knowledge, and the expansion and the increasing complexity of scientific research, widespread attention has been drawn to clearly visualizing the knowledge structure drown in massive amounts of literature as well as grasping development trends. Based on document representation learning and manifold learning algorithms, we suggest a method for constructing a semantic map (Smap). First, Doc2Vec is adopted to capture the high-dimensional semantic features between documents; then, UMAP (uniform manifold approximation and projection) is utilized to perform non-linear dimensionality reduction on the semantic proximity of documents. Finally, the kernel density estimation is employed to characterize the knowledge structure according to the heterogeneity of the document distribution. In the empirical experiments, we cover four scientific domains, ranging from thousands-level to millions-level of documents. Then, we construct an Smap, identify knowledge hierarchical structure, and analyze their dynamic evolution. Furthermore, using the classification system provided by Microsoft Academic Graph (MAG), citation relations, and keywords, we quantify the local purity of the document distribution on Smap and the correlation between the map distance and research distinction to verify the effectiveness of the proposed method. By comparing with controlled experiments, we further demonstrate the significance of the effectiveness of our method. This study expands the current methods of visualization systems in the scientific field and provides an alternative visualization method for scientific and technological information services. 2023 Vol. 42 (1): 74-89 [Abstract] ( 212 ) HTML (175 KB)  PDF (8874 KB)  ( 206 )

Intelligence Users and Behavior

90 Research on the Learning Engagement of Information Search Users: Based on Kolb's Learning Style and Cognitive Flexibility Theory Hot! Sun Xiaoning, Ji Fuchun, Liu Siqi DOI: 10.3772/j.issn.1000-0135.2023.01.008 The research scope of learning engagement, a popular topic in learning science, is increasingly growing. “Search as learning” generates a different understanding of interactive information retrieval as well as learning activities. Exploring the influencing factors of learning engagement can improve cognitive flexibility and knowledge transfer, develop the application and innovation ability of information search users, and provide a reference for optimizing the functional design of the information retrieval systems supporting learning objectives. In this study, the learning engagement of information search users was categorized according to three dimensions: behavioral engagement, cognitive engagement, and emotional engagement. The effects of learning style and cognitive flexibility on these three dimensions were explored via information retrieval experiment research. Data analysis and processing methods include search logs analysis, content analysis, self-reporting, analysis of variance (ANOVA), and the chi-square test. The results show that learning style has a significant impact on the learning cognitive engagement of information search users but no significant impact on their learning behavior engagement and learning affective engagement; cognitive flexibility affects the learning behavior engagement as well as learning emotional engagement of information search users but has no significant impact on learning cognitive engagement. Finally, learning style and cognitive flexibility only influence the three dimensions of the learning engagement of information search users, but display no interactive effects with them. 2023 Vol. 42 (1): 90-102 [Abstract] ( 152 ) HTML (180 KB)  PDF (1220 KB)  ( 192 )

103 Influence of Users' Search Path Characteristics on Information Search Effectiveness: An fsQCA Approach Hot! Zhao Yiming, Li Qian, Qiu Yumeng, Chen Yijin DOI: 10.3772/j.issn.1000-0135.2023.01.009 This study aims to explore how the characteristics of the search path affect search effectiveness. Through user experiment, two different search tasks were designed in the same task situation, and the four search path features—path length, depth, complexity, and novelty—were quantified and extracted. The influence of the combination of different path features on the search effectiveness was explored using fuzzy-set qualitative comparative analysis (fsQCA). Six search path feature configurations with high information search effect were found. Furthermore, three types of search path patterns with high information search effect were summarized—fast browsing, exploratory query, and conservative query types. This study is novel in that it revealed multiple concurrent causation among search path characteristics and search effectiveness from the configuration perspective. 2023 Vol. 42 (1): 103-112 [Abstract] ( 208 ) HTML (130 KB)  PDF (773 KB)  ( 345 )

Intelligence Discipline Development and Construction

113 Research on the Intelligence Work Model of Collaborative Technological Security and Technological Development Hot! Liu Mingyue, Yang Jianlin DOI: 10.3772/j.issn.1000-0135.2023.01.010 In the new era, scientific and technological intelligence work should not only pay attention to its security issues in the process of such development, but also to its development issues in the process of scientific and technological security maintenance, collaboratively discussing these issues. Based on the division of labor between security and science and technology development intelligence work, an intelligence work model for coordinating between these is proposed. This model integrates development thinking into scientific and technological security assurance, injects security maintenance awareness into the process of scientific and technological development, and realizes the mutual coordination of scientific and technological security and development information under the collaborative workflow of resource, technology, and task-oriented coordination. As such, it can identify and provide intelligence support for scientific and technological development opportunities in a security context as well as a perception of security risks within the scientific and technological development context. This model will ultimately serve the scientific and technological strategic decision-making of “overall development and security” to promote scientific and technological intelligence work that truly serves the needs of the national science and technology strategy. 2023 Vol. 42 (1): 113-126 [Abstract] ( 236 ) HTML (126 KB)  PDF (1759 KB)  ( 254 )