[Purpose/Significance] This paper explores how to better solve the entity extraction problem of low resource languages from a cross-lingual perspective.[Method/Process] With English as the source language,Spanish and Dutch as the target language,an unsupervised cross-lingual entity extraction method combining self-learning method and GRU-LSTM-CRF network is proposed based on the idea of transfer learning and deep learning.[Result/Conclusion] Compared with the supervised cross-lingual entity extraction method,the unsupervised cross-lingual entity extraction method proposed in this paper can achieve better results.In Spanish,the value of F1 is 0.6419,and in Dutch,the value of F1 is 0.6557.Cross-lingual knowledge is used to build a bridge between source language and target language to improve the effect of entity extraction of low-resource languages.

[Purpose/Significance] The purpose of this study is to identify the structure features by mining the texts of the planning and policies published in 10 provinces or cities,in order to disseminate key points in the texts,help refine the policies and planning,and inform the development of related industries.[Method/Process] The study examined 10 artificial intelligence(AI)policies and planning texts through content analysis and discourse analysis.[Result/Conclusion] The results revealed the similarities and differences in these texts.However,the policies and planning in different provinces and cities usually developed based on their own economic base,geographical characteristics,and cultural features,and thus there were many different features.Finally,the paper discussed some shortcomings in the planning and policies,based on which it gave suggestions for further refining the policies and planning texts.

[Purpose/Significance] Evaluating vital nodes rapidly in large-scale social networks is of great significance for the control of information dissemination.[Method/Process] In this paper,we proposed a local centrality vital node identification algorithm.The method comprehensively evaluated the K-shell of a node,its own clustering characteristics,the diffusion characteristics of its neighbors and propagation state of nodes,which simultaneously reflected the network location of the nodes,the topology of the neighbors and evolutionary features in time.The evaluation indicators were more comprehensive and efficient.[Result/Conclusion] The experimental results showed that the vital nodes identified by this method had a greater impact on the robustness of the network.After removing these nodes,the connectivity of the network was greatly affected,the network clustering coefficient was reduced,and the average path length was increased.Meanwhile,SIR model was used to evaluate the ability to spread nodes.Simulations of five real networks showed that our proposed method could improve the scope and average speed of information dissemination.

[Purpose/Significance] In order to seek for the endogenous influence of potential gradient and emotional conjugation on layered diffusion of public opinion,monitor and govern crisis information,and enhance the intention of sharing positive information.[Method/Process] The paper used VAR model,panel data model,and state space model to analyze the relationship between potential gradient,emotional conjugation and layered diffusion.[Result/Conclusion] The research found that(1)The impact of the potential gradient and emotional conjugation was not quickly transmitted to the layered diffusion,and presents a right skew distribution characteristic throughout the process.(2)The effect of potential gradient and the emotional conjugate was greater.(3)The change of the marginal influence of potential gradient and emotional conjugation on the layered diffusion showed an inverted U-shaped parabola,and the corresponding marginal influence effects were all relatively large.(4)The interaction effects between potential gradient,emotional conjugation,and layered diffusion were different among different demographic groups.

[Purpose/Significance] For the prosperity and development of an innovative platform,it is very important for users to have social interaction and contribute to the platform.In OIP environment,information interaction between users is frequent and the amount of information is increasing day by day.A large number of repeated and similar information makes information redundancy a problem that cannot be ignored gradually.Therefore,this paper studies the influence of content characteristics(information redundancy)in user interaction on user knowledge contribution.[Method/Process] The paper crawled the data of typical open innovation platform(LEGO IDEAS),used Python NLTK package to mine creative text content,and combined redundancy calculation method in information theory and zero inflated negative binomial regression model to make empirical analysis.[Result/Conclusion] The results showed that the amount of information acquired during user interaction positively affected the user's knowledge contribution and promoted users to publish more ideas.However,when there was redundant information between the acquired ideas,information redundancy would adjust the influence of the acquired amount of information on the user's knowledge contribution.In the process of user innovation,the content characteristics of user interaction are more important than quantity.Enterprises that build open innovation platforms should encourage users to actively participate in the interaction,and at the same time guide users to participate in the exchange and learning of low redundant ideas.

[Purpose/Significance] Virtual academic community has become an important platform for scientific research cooperation,but virtual scientific research cooperation groups are unstable.As stakeholders,all participants in scientific research cooperation expect to maximize their own interests,and there is a game relationship with virtual scientific research cooperation groups.The game analysis of scientific research cooperation behaviors in virtual academic communities is conducive to maintaining and promoting the formation of scientific research cooperation relationships in virtual academic communities.[Method/Process] Based on the hypothesis of rational person,this paper constructed KMRW reputation model to analyze the strategy choice and stability of research cooperation in virtual academic community,and introduced discount factor,incentive factor and punishment factor to further analyze the research cooperation.[Result/Conclusion] The research showed that in the research cooperation of virtual academic community,reputation would have a significant impact on the behavior of research cooperation individuals,while the existence of discount factors,incentive factors and punishment factors would increase the reputation value of research cooperation individuals,and promote the formation of virtual research cooperation groups to reach a balanced state,and accordingly put forward relevant strategies to promote the research cooperation of virtual academic community.

[Purpose/Significance] From the perspective of interaction,this paper analyzes the dynamic interrelations among core factors of knowledge collaboration in online encyclopedias,and reveals the collaborative mechanism of knowledge construction.[Method/Process] This paper analysed the basic process of knowledge collaboration in online encyclopedias based on co-evolution theory and proposed a knowledge collaboration model of online encyclopedias through the analysis of collaborative relationship among core factors.Taking Detective Conan as an example,this paper selected monthly time series data from 2007 to 2017,and uses vector error correction model,Granger causality test and impulse response function to analyze the dynamic interrelations among group size,cognitive conflict and content quality.[Result/Conclusion] The research results showed that there were significant dynamic interrelations among group size,cognitive conflict and content quality.Specifically,there was positive interaction between cognitive conflict and group size,as well as cognitive conflict and information quality.

[Purpose/Significance] With the gradual embedding of scenario elements in the industry university research community,the function of information sharing as a blood to promote collaborative innovation among the members of the industry university research community is becoming more and more obvious,which stimulates the new momentum of value creation of the industry university research community,and makes the information sharing of the industry university research community with scenario utility and adaptation.[Method/Process] Therefore,the information sharing scenario and information sharing scenario were included in the research of industry university research community information sharing.Starting from the connotation of industry university research community information sharing,the formation mechanism of industry university research community scenario was explored,and the realization path of industry university research community information sharing was designed.[Result/Conclusion] Based on the existing relevant literature,consulting experts in the industry university research community and experts in the information field to form an interview outline,using the semi-structured interview method to refine the information sharing adaptation strategies in three scenarios of "enterprise-university","university-scientific research institution" and "scientific research institution-enterprise".

[Objectives/Significance] To explore patent technology information mining methods applicable to enterprise technology innovation,and improve the utilization efficiency of patent information by enterprises.[Methods/Procedure] Deeply analyzed the connotation of enterprise technology innovation and the role and value of patent information in enterprise technology innovation.It started with the process and demand of enterprise technology innovation,which determined the content of patent technology information to be tapped in each stage of innovation,and thus selected technical information mining methods such as patent retrieval,patent quantitative analysis and qualitative analysis.[Results/Conclusion] A patent technology information analysis framework was proposed,which could basically meet the needs of enterprise technology innovation,and can provide reference for enterprises.

[Purpose/Significance] Regional government microblog can record and reveal some hot events closely related to the public.This paper builds the knowledge graph of regional government microblog to improve users' knowledge acquisition efficiency.[Method/Process] First,this paper captured contents of government microblogs and built a topic modeling using LDA.At the same time,we extracted the semantic triples from aforementioned contents by LTP.Second,the applied ontology of regional government microblog wes built to form the semantic structure of knowledge graph.Ultimately,Neo4j and D3.js were used to constructing,preserving and visualizing the knowledge graph.[Result/Conclusion] Based on the theoretical and practical research,we built knowledge graphs of regional government microblog,combing the semantic relations between organizations,figures and terms in microblogs' contents.This paper provides thoughts and methods for the construction of knowledge graph in social media field.

[Purpose/Significance] It is of great significance to study the network public opinion propagation of emergencies for improving public opinion guiding mechanism and promoting the modernization of national governance system and capacity.[Method/Process] Based on the dynamic model of infectious diseases,and the dual perspectives of heterogeneity and cross-evolution of public risk perception,this paper constructed the network public opinion propagation model by analyzing the cross-evolution mechanism of risk perception driven by emergencies.Besides,the basic reproduction number was also calculated.[Result/Conclusion] The empirical and simulation results of the network public opinion of "Beijing New Epidemic" showed that the smaller the heterogeneity of public risk perception and the greater the effect of emergencies was,the larger the scale and influence of network public opinion propagation were.The deterioration of emergencies would make the cross-evolution incline to the high risk perception,and intensify the spreading of network public opinion.However,the improvement of emergencies would promote the cross-evolution to the low risk perception,and it is helpful to control the scale of network public opinion.The government should start the emergency response immediately to reduce the level of public risk perception,realize the effective intervention of network public opinion,and maintain social harmony and stability.

[Purpose/Significance] In order to explore the dynamic mechanism to promote the evolution of Internet public opinion,this paper uses evolutionary game tools to clarify the internal relationship between the participants and the development of public opinion.[Method/Process] Firstly,the evolutionary game model of the incubation period,development period,outbreak period and recession period of Internet public opinion were constructed.By solving the game equilibrium and Jacobian matrix,the evolution path and associated conditions of Internet public opinion were analyzed,and the internal mechanism of the development and evolution of Internet public opinion was discussed based on the interest dependence between the players of the game.On this basis,on the MATLAB platform to "problem cable event" of Xi'an Metro Line 3 was simulated,and the effect of psychological satisfaction on game equilibrium was mainly tested.[Results/Conclusion] The key factors influencing the future trend of Internet public opinion in different development periods were also different;the main parties involved in the game jointly decided the direction of the evolution of Internet public opinion;the actions of government departments had a significant impact on the Internet public opinion;the number of psychological satisfaction factors was crucial to the decision-making of Internet users,and even could change the development trend of public opinion events.

[Purpose/Significance] It is of great significance to follow the trend of research hotspots in a subject with simple visualization in order to grasp the trend of research hotspots in a subject.Word frequency analysis is one of the hotspot analysis methods of subject research.Currently,there are many visualization tools based on word frequency analysis,but there are limitations that these visualization tools can clearly display the annual hot keywords in a concise visual form.[Method/Process] Therefore,this paper proposed a visualization method to measure the contribution rate of annual hot keywords to the total annual hot keywords by the ratio of the annual amount of articles in the subject area to the total amount of articles in the subject:based on the annual contribution rate and the 80/20 law,the threshold parameters were set and adjusted to control the number of annual high-frequency keywords presented,and the selected annual high-frequency keywords were sorted in order of word frequency and year to achieve visualization of research hotspots.[Result/Conclusion] Taking the field of "linked data" as an example to conduct empirical research,by analyzing the matching of the extracted high-frequency keywords with the existing high-frequency keyword extraction methods and comparing the visualization method presented in this paper with Citespace's co-occurrence knowledge map,Test and evaluate the feasibility of this method was.

[Purpose/Significance] To investigate the use of individual academic evaluation indicators in the field of picture situation in China,to provide reference for academic evaluation researchers,and to promote the improvement of individual academic evaluation system.[Methods/Process] Literature data on individual academic evaluation indexes of 9 CSSCI representative journals from 2014 to 2018 were collected,and content analysis method was used to investigate the application of indexes from four aspects:use scope,use environment,use motivation and use emotion.[Results/Conclusions] The analysis showed that the personal academic evaluation indicators in the field of picture and situation in China were in the development stage of the indicators mainly studied in category h index,while the traditional and simple bibliometric indicators,such as h index,publication volume and citation frequency,were mostly used.There are few studies on the integration of multiple indicators in individual academic comprehensive evaluation,so it is still necessary to continue to strengthen the research on academic evaluation,including the integration of multiple individual academic evaluation indicators and the further improvement of individual academic evaluation index system.

[Purpose/Significance] Through the text analysis and thinking of Suzhou Initiative on the Construction of First-class Academic Journals of Library Information and Archives Management Journal released by "The Construction of First-class Academic Journals in the New Era:The Second National Library Information and Archives Management Journal Summit Forum",this paper aims to provide reference and reference for library information and archives management journal to enter the publishing range of secondary literature journals and enhance the discourse power of library information and archives management journal.[Method/Process] This paper adopted the methods of content analysis and literature research to analyze and sort out the background and thinking of Suzhou Initiative on the Construction of First-class Academic Journals of Library Information and Archives Management Journal.[Results/Conclusion] Academic journals are an important platform to enhance academic discourse power.The secondary literature journals are the concentrated expression of academic discourse power,but the visibility of these journals in Xinhua Digest and Chinese Social Science Digest is not high.There are many kinds of validity to improve the discourse power of library information and archives management journal.

[Purposes/Significance] The research of information-encountering has become a hot topic in Library and Information Science and other disciplines,however,it is not easy to carry out empirical research due to the bottleneck of data sources and research methods.This paper aims to find out cues for future research through construction of knowledge map for information-encountering research.[Method/Process] With research papers published from 1995 till now at home and abroad,a knowledge map in the field of "information encountering" was formed by using the method of bibliometric and the tools of SATI,UCINET,NetDraw and CiteSpace.[Conclusion/Discussion] According to the results,the topic of information-encountering research at home and abroad has transformed from theoretical exploration to application research in new environment and with new technology from 1995 till now.As an interdisciplinary research field,the information-encountering research were carried out independently within the fields of Library and Information Science,computer related disciplines,Business Management,etc.Theories rooted in Library and Information Science research was spreading to related disciplines.The methods based on social network analysis and behavioral data analysis from computer related disciplines give insights into the research method of information-encountering in the filed of LIS.

[Purpose/Significance] This paper sorts out and summarizes the research on knowledge exchange effect in academic social networks to help academic colleagues grasp the development status,and provide ideas and directions for future research.[Method/Process] The researches were summarized into three aspects:research on influencing factors of knowledge exchange effect,research on evaluation method and evaluation index system of knowledge exchange effect,research on the strategy of improving the effect of knowledge exchange,and we analyzed the deficiencies of research results in each aspect.[Result/Conclusion] The existing researches had shortcomings such as single data source,confusion of similar concepts,incomplete consideration of constructing evaluation system,etc.Future research can make progress in these aspects.