[Purpose/Significance] Academic database is one of the infrastructures in scientific communication at present, and scientific communication entities such as researchers realize the creation of scientific value through interaction in the use of academic data.In order to clarify the characteristics and role of academic databases in the process of scientific value creation, [Method/Process] this paper attempted to apply the theory of value co-creation to compare and analyze the resource subscription model, the intelligent analysis value-added model and the open science service model of academic database from two dimensions: the interactive characteristics and value creation characteristics of knowledge service. [Result/Conclusion] The results show that there is no clear demarcation between these models, and they all realize the commercial value, academic experience value and scientific value in the co-creation process respectively, but there is still much room for exploration in the value co-creation mechanism, knowledge service experience, and scientific communication crisis and academic ethics.

[Purpose/Significance] The paper aims to provide local examples for expanding the discourse of inequality in information field from the perspective of algorithmic bias awareness. [Method/Process] This paper adopted Norman Fairclough's three-dimensional model of critical discourse analysis to analyze the discourse texts of algorithmic bias in social media from three dimensions: text, discourse practice and social practice. [Result/Conclusion] The research findings are stated as follows: ①Users' awareness of algorithmic bias comes from data and algorithm platforms, and the objects of bias include data, contents and users.Users have different emotions towards algorithmic bias; ②The interaction of discourse texts implies users' ideology; ③Users' awareness of algorithmic bias reflects the relationship of power game among stakeholders behind unbalanced distribution of information, inequality of information presentation, and digital inequality.

[Purpose/Significance] This paper proposes to combine text and image feature information to classify emergencies in a hybrid fusion way, aiming at the problems of single modal or modal absence and unsatisfactory classification effect in the research of emergency classification. [Method/Process] This paper proposed an Emergency Classification Model with Hybrid Fusion(ECMHF)based on multimodal hybrid fusion, which was mainly composed of three parts: text feature extraction classification model linked by BERT-BiLSTM, image feature extraction classification model based on VGG19, multimodal classifier composed of multimodal fusion layer that fused text and visual semantic features and prediction category output layer.Finally, the decision level outputs of all models were fused after weight allocation. [Result/Conclusion] The collected emergency news data(including natural disaster events, public health events, social security events and accident disaster events)were used as experimental samples for empirical research.The result shows that ECMHF model has the best neutral performance in all models, with the highest F1 value of 99.072%, 0.51% higher than the suboptimal model.ECMHF model with hybrid fusion strategy can effectively identify the types of emergencies.The test was only conducted on the collected data set, not verified in a more complex real environment.

[Purpose/Significance] As an external information resource of an enterprise, online comments are an important basis for users to make purchase decisions.By identifying the dynamic change of users' attention in online reviews, it can provide valuable reference for enterprises to make business decisions. [Method/Process] Based on signal theory and search products, LDA model was used to mine topics in online reviews, and then the intensity and changes of each topic in each time period were calculated and analyzed.Finally, Stata software was used to analyze the dynamic change of user attention. [Results/Conclusion] The research shows that the user's attention to the functional attributes in the product's internal signals has always been at the highest intensity; Attention to product and service signals has been on the rise; The attention to the external signal of the product shows a downward trend.

[Purpose/Significance] As a tool to guide users to browse and query on information platform, navigation plays a very important role in the efficiency and experience of information acquisition.This study takes the rumour-dispelling platform as a sample, aiming to explore the strategy of navigation function optimization from the perspective of user experience. [Method/Process] From the perspective of user experience, this paper adopted a combination of eye-tracking experiments and interviews to observe the differential effects of different navigation types, locations, information organization methods and other factors on user gaze indicators.Based on the theory of technological affordance, the navigation factors that affect user experience and usage intention were extracted, and then, some suggestions for navigation optimization were put forward. [Result/Conclusion] It found that the content delivered by navigation is the main factor affecting the user's gaze index, and classified navigation is more in line with user preferences.Interactive affordance, immersion affordance, searchability and information affordance should be the focus of the design and optimization of the rumor refutation platform navigation.

[Purpose/Significance] The limitation of granularity has not been broken through since a great deal of research on knowledge exchange were already emerged in perspective of citation data.This paper tries to explore relationship between knowledge elements from the granularity of text content, and then analyzes the role of each knowledge element in knowledge communication. [Method/Process] In this paper, knowledge elements were extracted from literatures by a method based on knowledge description rule, and their attributes such as topic and sentiment were calculated.Then, the relationship between knowledge elements was defined according to the characteristics of knowledge elements, such as topic, structure, sentiment and cited times.Finally, the relationship network of knowledge elements was constructed, and the role of a knowledge element in knowledge communication was identified according to the number and type of relationships connected with it. [Result/Conclusion] A total of 3187 knowledge elements are extracted from 207 papers in the field of"knowledge elements".This paper describes the distribution of knowledge element relationships and how the number and type of knowledge element relationship influences a knowledge element's role of knowledge communication.This paper realizes the semantic relation between documents from the granularity of knowledge elements, which deepens the research level of knowledge exchange from document carrier to knowledge content.

[Purpose/Significance] The paper is to sort out the research content of social media health information sharing behavior and look forward to the future research direction. [Method/Process] This paper systematically reviewed the literature on social media health information sharing behavior, analyzed related concepts, and used content analysis method to mine the time distribution, research methods, research focus and theoretical basis of existing literature, and constructed the content framework of social media health information sharing behavior research. [Result/Conclusion] Factors influencing sharing behavior, types of health information, functions of sharing media and characteristics of sharing groups are the research focuses of health information sharing behavior on social media.The shortcomings of the existing literature are reflected in the lack of in-depth research on the sharing behavior of social media attributes and distorted health information, the lack of innovation in research methods, the lack of rich theoretical sources, and the urgent need to further expand the research situation.

[Purpose/Significance] From the three dimensions of personal, environment and technology, the paper analyzes the influencing factors of WeChat Subscriptions users' unfollow intention, and provides references for the WeChat Subscriptions platforms to improve users' follow rate. [Method/Process] Based on the personal-environment-technology(PET)framework, a theoretical model of WeChat Subscriptions users' unfollow intention was constructed.Through the network questionnaire survey, 346 valid questionnaires were collected, and the structural equation modelling(SEM)and the fuzzy set qualitative comparative analysis method(fsQCA)were used to analyze the data. [Result/Conclusion] ① Path analysis shows that: interest shift, expectation disconfirmation, task-technology misfit, uncertainty, the number of WeChat subscription follow, and alternative attractiveness all positively affect users' unfollow intention; ②Configuration analysis shows that there are three antecedent configurations that affect users' unfollow intention of WeChat Subscriptions.

[Purposes/Significance] To address the problem of how to effectively improve the accuracy of technological innovation partnership prediction, fusion prediction indicators based on network topology indicators and technology subject attribute characteristics indicators are proposed to be constructed. [Method/Process] Starting from the characteristics of technological innovation cooperation relationship, the technological factors affecting the formation of technological innovation cooperation relationship between subjects were taken as important factors in the prediction of cooperation relationship, and the corresponding characteristic indicators of technological subject attributes were extracted from two aspects of technological similarity and technological complementarity.By fusing with six network topology indicators respectively, class Ⅰ, Ⅱ and Ⅲ fusion indicators were constructed.In order to verify the effectiveness of these indicators, the field of marine engineering equipment was chosen to be the research object. [Result/Conclusion] Fusion indicators can effectively improve the prediction accuracy of the technological innovation partnership of marine engineering equipment, with the highest increase of 29.23% and the average increase of 14.99%, among which the average increase of Class Ⅰ and Class Ⅱ fusion indicators are around 10% and the average increase of Class Ⅲ fusion indicators is 24.12%.

[Purpose/Significance] The article attempts to solve the problems of the complex structure, inadequate teachers, and limited time and space of inclusive digital literacy education by exploring modules in the frame, specific risks, and countermeasures for embedding AI generated content into digital literacy education. [Method/Procedure] The article pointed out the integrating digitization and intelligence in the digital ecosystem with the high-quality development driven by innovation has made digital literacy a survival skill for individuals.It indicated the technical deconstruction and application value of a large-scale and cross-modal artificial intelligent algorithm model, described modules and risks countermeasures of embedding AI generated content into digital literacy education. [Result/Conclusion] The ordered optimization of embedding AI generated content into digital literacy education will provide strong support for building a universal digital literacy education mechanism and helping to achieve equitable development in the digital society.

[Purpose/Significance] Analyzing the topics distribution of academic integrity policies and literatures, and finding the correlation and trend development between them can provide a basis for policy formulation and can also provide a reference for the direction of thesis research. [Method/Process] The article collected academic integrity policy and literature data, and used LDA topic clustering and similarity analysis methods to find hot topics and analyze the correlation between literatures and policies. [Result/Conclusion] Through the method of"LDA-TF-IDF-Sankey chart"uncover the relationship between academic integrity policies and research hotspots, which can support the methodology for similar research.Research showed, the development of literatures related to academic integrity has obvious lag in the development of academic integrity policies, but with the development of the time and the progress of research, the lag is weakening.

[Purpose/Significance] The research on the situation, gaps and uses of open government data(OGD)users' participation process is of great significance to increase users' engagement and release the value of OGD. [Method/Process] Based on the perspective of user cognition and the sense-making theory, an process framework of three factors(situation-gaps-uses)was constructed to analyze the process.And the OGD users' participation was specifically divided into five situations in this study.Besides, this study investigated the gaps of each situation users meet and the ways users use to cross the gaps through the micro-moment time-line interview which helped restructure the process. [Result/Conclusion] This research finds that there are different gaps in each users' participation situation.And these gaps are about three aspect problems: participation environment, platform construction and data quality.As for these problems, specific suggestions are put on how to create better environment and improve the basic foundation for further OGD user participation.

[Purpose/Significance] Activate government data resources, build a solid foundation for data storage, and give full play to the role of government data elements as the engine of economic and social development. [Method/Process] The papaer built a blockchain-based dual-chain storage model for the process of open sharing of government data, expanded the storage space of the blockchain system through "off-chain expansion", and carried out a case application study of the model based on the empirical analysis of the model in the "Quancheng Chain"of Jinan City, and summarized the reference and inspiration of the research on the data storage of other provinces and municipal governments. [Result/Conclusion] The model innovates the data storage method with the operating mechanism of "on-chain service, off-chain storage", which guarantees the safe sharing of data while taking into account the storage needs of massive amounts of data, and helps to improve the open government data sharing system, completely unleash the potential of data elements, and provide references and ideas for data storage in other provinces and municipalities.

[Purpose/Significance] Data Governance is one of the core foundational task of digital government construction.The summary of the construction experience of government data governance system in European Union(EU)can provide useful reference for Chinese digital government practice. [Method/Process] Based on the literature review of data governance system, this paper analyzed the strategic vision, supportive measures and technical tools of the EU government data sharing and open data. [Result/Conclusion] The results show that the EU has already established a set of government data governance system with the vision of building a single unified data space.Besides, the system is consisted of mutually supportive policies and regulations, collaborative governance structure and interoperability framework.The construction experience has many implications to data governance for realization of government data value in China.

[Purpose/Significance] To deepen the application of blockchain technology in network public opinion, promote the construction and improvement of the network public opinion traceability system, and improve the scientific and predictable risk management of sudden public opinion. [Method/Process] Based on situation awareness theory and blockchain technology, this paper built a multi-social platform network public opinion traceability system, adopted the dual storage mechanism of blockchain and IPFS for mass storage of public opinion data, used smart contracts for monitoring and early warning of risky public opinion, and realized accurate traceability of network public opinion based on content identifiers CID and event ID. [Result/Conclusion] The network public opinion traceability system for public health emergencies from the perspective of blockchain has been improved, the traceability and predictability of public opinion risks among multiple platforms have been improved, and accurate public opinion early warning has been provided for relevant government regulatory departments and decision support.

[Purpose/Significance] Network public opinion risk information involves many elements.In order to improve the perception effect of network public opinion risk information, a network public opinion risk information perception model of environmental emergencies is constructed in the new media environment. [Method/Process] The risk index set of network public opinion information flow of environmental emergencies was established, the evaluation index weight was given, and the weight was discretized to form a similarity coefficient matrix.This paper analyzed the evolution law of the core elements of network public opinion communication through three parts: the modeling of public opinion subject, the establishment of public opinion information model and the establishment of Internet users' emotion model.On this basis, the propagation probability was calculated to determine the involvement degree of users in disseminating information, and the risk of information was perceived by logistic regression model.The decision tree was used for public opinion risk early warning to realize the network public opinion risk information perception of environmental emergencies. [Result/Conclusion] The experimental results show that the analysis accuracy of the event public opinion emotion evolution after the application of the constructed model is higher, the perception accuracy of the number of forwarding and malicious comments of each event is higher, and the early warning effect is better, which effectively improves the perception effect of network public opinion risk information of environmental emergencies.

[Purpose/Significance] The study of the Allee effect of public opinion guide is helpful to enrich the scientific exploration of public opinion guide, and provide theoretical support for solving the real public opinion guide problems. [Method/Process] This paper analyzed the Allee effect of public opinion guide, and built the interaction model of public opinion conductor under the action of Allee effect based on logistic model.Through simulation, it explored the evolution characteristics of the guiding effectiveness of public opinion conductor under the action of Allee effect, and clarified the mechanism of Allee effect on ideal guidance and general guidance of public opinion conductor under the reciprocity, competition and purging interaction. [Result/Conclusion] Through theoretical analysis and simulation research, the influence of Allee effect on network public opinion guidance is verified.