[Purpose/Significance]This paper aimed to improve the performance of entity recognition from Chinese electronic medical records by integrating BiLSTM-CRF neural network and dictionary resources with prior knowledge of the field.[Methods/Process]The BiLSTM-CRF neural network model was used and experimented on the desensitized Chinese medical record data provided by CCKS-2017 evaluation task.The Word2Vec word embedding was combined with external dictionary in the neural network to improve the performance of entity recognition.[Results/Conclusions]The results showed that compared with the traditional CRF and the pure BiLSTM-CRF model,adding dictionary resources with prior knowledge could achieve better performance of entity recognition,with the highest F1 of 90.41%.BiLSTM-CRF could improve the performance of entity recognition over CRF,and prior dictionary knowledge could further enhance the performance of neural networks.

[Purpose/Significance]The research aimed to discover the focus of city identity in a period of time,and to analyze the public's overall impression of city and its dynamic changes.[Method/Process]Firstly,this paper proposed evaluation index,data source and calculation method for discovery of the focus of city identity.Secondly,it normalized the evaluation index by min-max normalization,and established evaluation model to calculate different cities' identities by linear weighted sum method.Finally,it discovered the focus of city identity of the six central China provincial capitals in different periods,based on the Q&A data from "Zhihu" platform.[Result/Conclusion]It was operable to use the method in this paper to discover the focus of city identity in a period of time from the massive Q&A data.

[Purpose/Significance]The core of the research and judgment work of network public opinion on the emergency initiated by the major epidemic is to predict the evolution rules and development trends of the subjects embedded in the multi-scale spatiotemporal social network in the Internet context.Normalized public opinion research and judgment often rely on the accumulation of experience and use fuzzy inference.There is a black box for the analysis of sudden public opinion dynamics involving the major epidemic.[Method/Process]Under the guidance of the theory of spatio-temporal structure,this paper takes spatio-temporal big data as an important tool for this type of network public opinion research,incorporates spatio-temporal characteristics into the observation variables of the development of network public opinion on the emergencies triggered by the major epidemic,and explores the matching rules of of big data sources and massive public opinion information in spatio-temporal dimensions.This paper seeks the integration path between space-time data modeling and research and analysis of public opinion,combing the dynamic global research process of network public opinion on the emergencies initiated by the major epidemic,and constructs a multi-dimensional polymorphic spatio-temporal big data analysis framework for researching and judging.[Result/Conclusion]Provide government practice departments with a targeted,qualitative and quantitative combination of scientific and accurate research and judgment system,and provide ideas and decision support for the public opinion crisis management of public health events in China.

[Purpose/Significance]Introducing gamification into exploratory search system can stimulate users' interests,encourage continual using,support interaction,and improve users' utilization of exploratory search systems.[Method/Process]Based on the gamification framework and the operation mechanism of exploratory search system,game elements were combined,game modules with specific functions were designed and introduced into the system interface to realize the application of game elements in the non-game scenery exploratory search.[Result/Conclusion]A conceptual model of gamification exploratory search system composed of game layer,user layer and system layer was constructed.Twenty game elements were composed into 11 game modules,including 5 independent modules and 6 composite modules.These game modules were embedded in 13 pages designed by Axure RP,including gamification exploratory search interfaces and gamification Interactive interfaces.The conceptual model and prototype design proposed in this paper can provide theoretical support for the development of gamification exploratory search system.

[Purpose/Significance]In view of the rich semantic information of energy policy,this paper studies the effect of convolutional neural network model on energy policy text feature classification and recognition under different environments,and proposes optimization methods to assist the automatic classification operation of energy policy information resources,so as to facilitate researchers to better interpret energy policy.[Method/Process]This paper used character-level and word-level convolution neural network model to test the effect of automatic classification and recognition of energy policy texts in different environments.It made a comprehensive comparative analysis from the perspective of title,content and core topic sentences,and extracted different proportion of core topic sentences by Doc2Vec for optimization experiments.[Result/Conclusion]With the increase of the extraction rate of core topic sentences,the average value of F1 was normal distribution.When the extraction rate is 70%,the balance was reached.The average value of F1 evaluated by the neural network model was 83.45%.Compared with other methods in the experiment,the using of topic information which was extracted with Doc2Vec effectively improved the automatic classification effect of the convolutional neural network model.

[Purpose/Significance]With the trend of mobile reading panoramic ecology development,there are new theoretical and practical significances in research on the motivation and configuration path of user adoption.[Method/Process]Based on the theory of TAM model,this paper constructed a conceptual model of mobile reading user adoption,closely combining with literature research and new features of mobile reading.Qualitative Comparative Analysis of fuzzy sets and multivalued sets(QCA)were used to analyze asymmetric multiple conjunctural causation among variables.[Result/Conclusion]It was found that:1)Usefulness was the main factor for adoption or rejection,Interaction was the main factor for adoption but secondary factor for rejection;2)Usability,Playfulness,Mobility and Cost were the main factors for rejection but secondary factors for adoption;3)there were two paths to promote adoption:one was good Usefulness,good Usability,good Playfulness,good Mobility and good Interaction,the other was good Usefulness,good Mobility,good Interaction and satisfied Cost;4)To prevent rejection,we should focus on the negative effects of a single factor or a combination of all factors except poor Interaction.

[Purpose/Significance]By analyzing the information sharing mode of the online community users,the information sharing characteristics of the online community are summarized,which will help to strengthen the cohesiveness and participation of users in the community and provide guidance for the benign development of the online community platform.[Method/Process]Based on the information niche theory,the star network community and the industry vertical network community were selected as representatives,using case analysis and comparative analysis methods to analyze the different modes in practical applications and compared their similarities and differences.[Result/Conclusion]The research in this paper showed that these four typical information sharing modes in the online community can help users to achieve efficient and comprehensive information exchange and communication.The model can help them provide more kinds of information services for users and create a good information atmosphere.

[Purpose/Significance]This paper studied the current development status and type characteristics of academic social networks,and compared and analyzed academic social networks at home and abroad,aiming to provide reference for the future development of academic social networks in China.[Method/Process]Using literature research method,network research method,case analysis method,comparative research method and other methods,the current development status of academic social network was investigated and classified,and different types of academic social networks and academic social networks at home and abroad were distributed in disciplines and target users.The characteristics of interface,function application,operation mode and other aspects were analyzed and compared.[Result/Conclusion]From the aspects of market positioning and functional innovation,data integration and operation concept,stakeholder and cooperation,content quality and user stickiness,this paper proposed enlightenment and suggestions for China's academic social network.

[Objectives/Significance]Competitive intelligence is an important tool for enterprises to formulate scientific and technical planning,which has an important support for enterprises to implement the strategy of technological innovation.However,due to the constraints of financial and human resources,enterprises are unable to meet their needs of competitive intelligence based on their own strength,and there is a demand for competitive intelligence services provided by external organizations.[Methods/Procedure]From the aspects of the type of service subject,the content of service and the way of service,this paper carried out the research on the subject of competitive intelligence service oriented to the enterprise science and technology planning.[Results/Conclusion]It was suggested that enterprises made diversified competitive intelligence acquisition strategies to promote the effective development of enterprises' scientific and technological planning.

[Purpose/Significance]This paper aimed to explore the influencing factors of ability of screening health information of Chinese 4 urban residents,with the expectation of providing theoretical reference for improving realistic health information service and enhancing the ability of screening health information,and then of supporting Health China construction.[Method/Process]The related data of ability of screening health information from 37 urban residents were collected using semi-structured interviews,11 subcategories of influencing ability of screening health information of Chinese urban residents were coded and 4 main categories were focused based on Grounded Theory.Based on the above analysis,the model of influcencing factors of ability of screening health information was constructed.[Result/Conclusion]Information factor and social class influenced ability of screening health information directly,social class and individual characteristics factor and system factor separately moderated the influences of information factor on ability of screening health information.

[Purpose/Significance]This paper aims to find out the attitudes of network users towards the use and protection of personal data through the survey on the willingness of Chinese network users to control their personal data,and to seek ways to protect the personal data of network users.[Method/Process]This paper used questionnaire survey to obtain the tendency of Chinese network users to use private data.Using literature research and empirical methods to show the existing views and the reality of personal data authorization intention.[Result/Conclusion]The results showed that only about 30% of Chinese interviewees were willing to disclose their personal data when they received free services or discounts.The vast majority of Internet users were aware of the potential danger of personal data and hoped to protect it through technology,policies,regulations and industry self-discipline.The research of this paper would help to understand the willingness of personal data control and emphasize the importance of China's personal data protection legislation.

[Purpose/Significance]Analyzing the current status of B2C online merchants' privacy policy and the policy differences among the different categories of online merchants,and concluding the practical inspiration.[Method/Process]Using content analysis method,the privacy policies of the selected 30 B2C websites,which were divided into five categories,were studied from two aspects:external features and content features.[Result/Conclusion]All online merchants had privacy policy,but there were still problems in visibility,readability and comprehensiveness of privacy policy.The privacy policies of every category did not differ greatly.The significant differences were between the group purchase websites and the vertical e-commerce websites,the group purchase websites and the online banking payment websites.

[Purpose/Significance]There is an increasing conflict that local governments emphasize on investment but neglect risks management during Open Government Data(OGD).The dilemma of lags both in academic and practical domains makes the issue more serious.[Method/Process]By constructing a "process" analytical framework,this paper aimed to identify risk factors of OGD.[Result/Conclusion]The result showed that lack of publicity of data demand,imitation effect of policy making,uncertain open standards,inefficient data decentralization,low motivation for citizen participation,non-awareness of data management,technical agency trap and weak data competency composed risk factors of OGD.Additionally,these risk factors could be classified into Contextual,Valuable,Instrumental and Mixed Risks.Risks prevention and control systems should be established,and the multiple potential values of OGD could be further released.

[Purpose/Significance]In general,the linear multi-attribute evaluation value does not have the direct judgment function of evaluating the superiority and inferiority,it cannot judge whether it is good or not based on the evaluation score.This problem is an implicit problem,which is particularly prominent in the evaluation of science and technology,and has not received enough attention.[Method/Process]Based on the analysis of the influence mechanism of multi-attribute evaluation value evaluation function,this paper proposed a dimensionless processing based on sigmoid function,analyzed its influence on natural weight and linear science and technology evaluation,and conducted an empirical study based on JCR2017 economics journal.[Result/Conclusion]The results showed that it was not proper to judge the quality of the evaluation object directly according to the evaluation value.The judgment function of evaluation value depended on the evaluation index dimensionless method and evaluation method.Sigmoid dimensionless could effectively eliminate the natural weight problem.Different dimensionless methods will affect the results of linear evaluation.The use of maximum dimensionless for bibliometrics questionable questionable.It was suggested to popularize the sigmoid function dimensionless in the science and technology evaluation.

[Purpose/Significance]Exploring and analyzing the spatial and temporal distribution and evolution characteristics of domestic journal articles by using exploratory spatial data analysis method,the depth and breadth of spatial measurement dimension research in Library and information field are expanded.[Method/Process]This paper analyzed the spatial pattern of the distribution and evolution of journal paper output in 31 regions of China from 2008 to 2017,by taking the journal paper output of various regions in China as the research object,taking the number of journal published by scientific researchers in CNKI as the measurement index,and using the method of exploratory spatial data analysis.[Result/Conclusion]The results showed that the output of journal paper was more in the east than in the west.The absolute difference of journal paper output among regions increased,while the relative difference decreased.The agglomeration of journal paper output showed a trend of rising first and then declining.The intra-regional and inter-regional differences of the output of journal papers in the three regions were narrowing.The polarization phenomenon of regional journal paper output was weakened and the spatial agglomeration pattern was relatively stable.

[Purpose/Significance]In order to explore the factors influencing the citation and expand the international influence of Chinese papers in Humanities and Social Sciences.[Method/process]Base on the samples of Chinese papers in the field of Humanities and Social Sciences which were collected by WOS in 2014,this paper used statistical analysis methods such as non-parametric test,multiple linear regressions to analyse the relationship between the characteristics of paper and citation.[Result/conclusion]1)The results showed that 25 factors except page were significantly associated with citation in univariate analysis.2)Results Multivariate analysis showed that co-authored obtained more citation than sigle-author in author-institution indicators;Among the journal indicators, IF(2),Q1,Q2,Q3 journal(based on Q4 journal),Elsevier(based on University publishing platform)had a positive effect on paper citation;Among the literature characteristics indicators,english language,open access and references had a positive effect on citation,while conference papers(based on review)had a negative effect on citation.Altmetrics such as usage can be regarded as the potential manifestation of paper citation.

[Purpose/Significance]Visualization is a general technical requirement and a trend of information services in the new situation,and better utilizing of this technology depends on investigation and evaluation of its utility.[Method/Process]From aspects of theory employing,assessing ways,task,dimension and index of evaluation by reviewing relevant researches and practices.[Result/Conclusion]The methods used in the assessment of visualized applications were quite diverse,but most of the evaluations were mainly based on experimental methods.The utility of the assessment was mostly about cognition.Emotional and social utility were rarely discussed and measured.There was a lack of an assessment generic framework and standard test set(of task,data)for evaluation.Corresponding to this,there were many opportunities for innovative practices in the evaluation of information visualization applications.

[Purpose/Significance]This article systematically reviewed the research literature on dynamic ontology construction at home and abroad,which aimed to provide reference for further study and development of dynamic ontology.[Method/Process]This article adopted methods such as synthesis,induction,and comparison.It included four parts:the definition of dynamic ontology,knowledge modeling,construction process and evaluation standard.[Result/Conclusion]The results showed that related areas made a lot of achievements,but there were some problems.The next step in domestic research in this area can be:focus on text-oriented dynamic ontology construction method,develop extended functions and pay attention to the sharing of dynamic ontology,complete the evolution of dynamic ontology by publishing content and behavior according to users,target research on dynamic ontology construction standard.

[Purpose/Significance]By collecting relevant researches of domestic and foreign scholars on social network user behavior,the research frontier and development trend of social network user behavior at home and abroad are deeply understood.[Method/Process]Literature analysis and knowledge map visualization were used to summarize the origin,development and research hotspots of social network user behavior at home and abroad,and to analyze the development trend of future research on social network user behavior.[Result/Conclusion]Research hotspot analysis results showed that foreign scholars mainly focused on the continuous use behavior of social network users,the impact of privacy risk on user behavior,and the fatigue behavior of social network users,while domestic scholars mainly studied the information interaction and forwarding behavior,use behavior and user behavior characteristics of social network users.The future development trend of social network user behavior will focus on user behavior portrait,user privacy,social network special user group and social network burnout behavior.