[Purpose/Significance] Emoticon has been an important element in current social media information exchange.The study of its usage differences and semantics in different contexts has certain significance for informal information exchange research.[Method/Process] In this paper,we used the actual session data of WeChat groups as the data source,and designed the emoticon extractor to obtain the emoticon usage data in the WeChat group session sample.Based on this,we studied the usage behavior of emoticon in information exchange process,focusing on usage differences in different types of WeChat groups and groups of different ages and semantics of the emoticon in different contexts.The usage rule of emoticons was also explored in this thesis.[Result/Conclusion] We found that in the process of WeChat group information exchange,the message type structure of different types of WeChat groups was quite different;the usage behavior of emoticons of different types and ages was quite different;the emoticons presented different semantics under different dialogue samples;the use of emoticons in WeChat groups was consistent with the Power-Law distribution,and so on.

[Purpose/Significance]To construct and identify the 3-D structure of the influencing factors of information security risks in smart cities,it is helpful to improve the ability of information security risk response in smart cities.[Method/Process]Combined with the reality of smart city management,based on the three dimensions of environment,logic and organization,a 3-D structure of influencing factors of information security risk in smart cities was constructed.Then the system of influencing factors was constructed by Delphi method,and the influencing factors were correlated by DEMATEL method analysis.[Result/Conclusion]According to the index of each influencing factor and its ranking,combined with the causal relationship map,the influencing factors of information security risk in smart cities were identified.

[Purpose/Significance] The rapid development of social media websites has contributed a lot of data to us.Through further mining of these data,we can realized the personalized recommendation's service.[Method/Process]This paper used the metadata information in the geo-tagged dataset,combined with the DBSCAN clustering algorithm based on density and the TF-IDF statistical method to extract and identify the local spots,and then calculated the hot of the local spots in combination with the seasons.At last,we used Hybrid filtering algorithm to achieve personalized travel recommend services.[Result/Conclusion]The Geo-tagged dataset from the Flickr website proved the effectiveness of the paper's extraction method.

[Purpose/Significance]Outlier detection from mass fundamental data could provide important information about latent terrorists for early warning of counter-terrorism.[Method/Process]Abnormal people that differed from most data objects must be identified by using clustering method to classify mass terror related data with composite properties.Additionally,abnormal people acted as the basic data source to find terror related people.These subjects who were excavated out had the larger degree of similarity with terrorists.[Result/Conclusion]This method provided an idea to detect high sensitivity people from multiple data streams.It could accelerate the speed of handling counter-terrorism intelligence and provide reference for counter-terrorism disposition by means of swiftly finding the terror related people.

[Purpose/Significance]Online reviews play an increasingly significant role in solving information asymmetry in the decision-making process of consumer online shopping.In order to optimize the user experience of consumers,it is necessary to timely and correctly dig out the online review information that is useful to consumers.Therefore,it is of great significance to explore the factors influencing the usefulness of online reviews.[Method/Process]Based on signaling theory,this paper constructed a model of factors influencing online review usefulness from the signals related to the review content,reviewers and the review feedback,in which the moderating role of the product type was considered.On the basis of previous researches,the review readers' feedback to the review and the signal environment influence were explored.[Result/Conclusion]The objective online review data were obtained through Amazon China website,and the factors influencing online review usefulness were empirically tested.The results showed that negative reviews,reviews with more number of words,reviews with pictures,reviews with the reviewer disclosure,reviews produced by the reviewer with higher rank were more useful for consumers.The product type had a moderating role in review sentiment and review pictures influencing review usefulness,moreover,the influence of signal environment was verified.

[Purpose/Significance]With the view to explore the influence mechanism of cyberspace field situation on opinion bias in crisis event,classify and grade differently characterized situation in the network,monitor and govern network opinion,and enhance the communication effect of positive opinion from the perspective of field situation.[Method/Process]With context relation,topic interaction and structure heterogeneity as independent variables,opinion bias as dependent variable,cognition convergence and emotion coupling as mediating variables,the paper constructed influence mechanism model of effect of cyberspace field situation on opinion bias in crisis event.[Method/Process]And the sample data was obtained through questionnaire survey.The data was analyzed using research methods SEM,and were processed through AMOS22.0.[Result/Conclusion]And it concluded that context relation,topic interaction and structure heterogeneity had significant effect on opinion bias through mediating variables cognition convergence and emotion coupling.Among these,the effect of structure heterogeneity was the most intensity,topic interaction more intensity,and context relation weakest intensity.Besides,there was differential effect of field situation on opinion bias between-and in-gender,age and education.And finally,the paper analyzed and discussed the research results,and indicated the theoretical and practical implication of the research conclusion.

[Purpose/Significance]Exploring the influencing factors of college students' information avoidance behavior in social networks can not only provide some theoretical support for the society to deeply understand the information avoidance behavior of college students in social networks,but also provide some reference and reference for social platform to optimize user information avoidance behavior.[Method/Process]Through semi-structured in-depth interviews,the data of 30 college students' information avoidance behaviors in social networks were collected.The NVivo 11 software was used to encode and analyze the original data,and the 10 basic categories that affected the information avoidance behavior of college students in social networks were summarized and focused on three main categories,it constructed a theoretical model of college students' information avoidance behavior in social networks on this base.[Result/Conclusion]The study found that cognitive factors were the direct influencing factors and internal drivers of college students' health information avoidance behavior in social networks.Information factors and social factors were indirect factors and external context factors.Among them,cognitive factors were the most vital influencing factors of college students' information avoidance behavior in social networks.

[Purpose/Significance]This paper investigated the needs and preferences of users of knowledge-based payment app and the existing problems.Aiming at the existing problems,it constructed a knowledge-based payment app service system,with a view to providing some reference for the development of knowledge-based payment app service construction.[Method/Process]It used the questionnaire to conduct online research on user needs,user preferences,and knowledge payment APP problems,and analyzed the research results.[Result/Conclusion]On the basis of investigation and analysis,the framework and implementation strategy of knowledge payment app service system were proposed based on user's needs.

[Purpose/Significance]Government affairs information release as an important part of government affairs service,the research on its quality evaluation is helpful to find out the shortage of work and to provide guidance for its service improvement.[Method/Process]This paper combed the domestic and foreign government affairs information release research and the practice,in the analysis of the PZB gap theory and its derivative model SERVQUAL,combined the actual situation of China's government affairs information release work,innovated the evaluation dimension and index.The evaluation model of service quality of government information affairs information release was constructed.[Result/Conclusion]Based on the empirical research for the public,this paper clarified the public perception of the service quality of government affairs information release,that is,the current service quality had not met the public expectations,and should be improved from the aspects of release platform construction,political and civil interaction skills,information content service and publicity and promotion.

[Purpose/Significance]In order to help the tourist attractions to shoot excellent virtual promotional video of scenic spots and to provide high-quality travel decision-making information for travellers.[Method/Process]This paper analyzed 22 typical virtual promotional video of scenic spots by content analysis method.[Result/Conclusion]The results showed that the features of these virtual promotional video of scenic spots include "mainly focus on natural elements" "mostly adopt shooting techniques of mobile,upward or horizontal angle of view" "all use background music while some of them additionally use background sound" and "interactive elements are introduced in a minority of them".There were some suggestions that should be paid attention to when shooting virtual promotional videos of scenic spots.The length of the promotional video should be controlled from 1 to 3 minutes to effectively display the most distinctive features of the scenic spot.Suitable shooting method should be chosen for the characteristics of scenic spots to provide travelers with immersive experience.Humanistic elements should be added to provide travellers with rich and vivid travel decision-making information.Background sound and background music that matched the picture should be used to provide a pleasant audiovisual experience for travellers.The quantity of interactive elements into virtual promotional video of scenic spots should be increased to provide a sense of presence for travelers.

[Purpose/Significance]According to the concurrent Internet circumstance,the paper proposed a method to classify certain internet public sentiment instance based on probability analysis.Therefore it provided knowledge which was need for decision making in Internet public sentiment crisis-response.[Method/Process]Using belief network,it constructed the Internet public sentiment index system and Internet public sentiment case matching model.[Result/Conclusion]Using 43 out of 48 Internet public sentiment crises as training data,trained the belief network.The remaining 5 crises were then used as test set.All 5 crises were labeled correct according to the belief network.[Conclusion]The Internet public sentiment index system and belief network could classify Internet public sentiment crisis correctly,therefore providing solid basis for decision making in public sentiment crisis-response.

[Purpose/Significance]The formation of the Internet society makes universities in the Datasphere.The data search habits of university users are different.Web search is still the basic way to access user information.Due to the unstructured big data,large amount of data,data mining and language processing,users are not satisfied with the search results.The "small data" of university users can reflect their explicit demand,and the data analysis and mining can reflect their implicit demand.User-centered service system is one of the most effective methods to improve service efficiency,user satisfaction and loyalty.[Method/Process]On the basis of literature analysis,this paper analyzed the composition,collection significance,norms,methods and USES of college library business data.This paper explored the methods of data collection in university libraries and the necessity of constructing "Datasphere" to support the development of intelligent services in libraries.[Results/Conclusion]The formation of "Datasphere" in university libraries was the inevitable way for smart libraries to accompany the development of Internet society,and also the basis for the effective implementation of smart services.The normalization and standardization of "data circle" was also an important issue in the field of smart library.

[Objective]In order to study social networks and study the transmission process of public opinion between Weibo network and WeChat network.[Method]Using the scale-free directed network and BA network to simulate the microblog network and WeChat network environment respectively,the coupled network carrier was designed by the characteristic connection relationship.Based on the SEIR model,the user's communication psychology was fully analyzed.At the same time,the individual had the characteristics of interest attenuation and memory effect,and constructs a social network public opinion propagation model based on coupled network.[Results]The experimental results showed that the constructed two-layer social network public opinion communication model could better reflect the public opinion communication process in real life.The interaction between users in multiple layers of social networks accelerated the flow of public opinion information and expanded the lyrics.Impact,the control of the inter-layer propagation threshold was the key to managing the lyrics of multi-layered social networks.

[Purpose/Significance]Mastering and understanding the evolution law of emotions in university under the microblog environment.It is great significance for relevant departments to strengthen the monitoring and supervision of university,so that universities can take measures to deal with the vicious transmission of negative public opinion events.[Method/Process]This paper analyzed text features through text mining and using word cloud visualization.This paper classified the online user comment texts based on the naive Bayes classifier.This paper combined the evolution of user emotions and the development cycle of public opinion events to dynamically display the emotional evolution map of university.[Result/Conclusion]The proportion of negative emotions of netizens reached its peak during the spread of lyrics.The proportion of neutral emotions was the lowest in the period of lyric spread,and the proportion of positive emotions barely changed during the lyric cycle.Through the research on the emotional evolution map of university in microblog environment,it provided new theoretical support for the research of university public opinion in microblog environment.At the practical level,it played a targeted role for the public opinion supervision department to timely monitor and effectively guide the public opinion of university.

[Purpose/Meaning]In the era of big data,with the gradual deepening of public data,the role of the government in the process of data opening is increasingly important.[Method/Process]The article started from the status quo of the openness of public data,proposed the main responsibility of the government at the theoretical level and the optimization of the specific duties at the practical level.[Results/Conclusions]In the process of public data opening,the optimization ideas of the main responsibility of the government mainly included broadening data sources,strengthening platform construction,eliminating data island,studying the commercial development of open data,constructing demand-driven data disclosure,implementing "innovation-driven development" with big data and implementing pilot system of data opening.

[Purpose/Significance]Current personal information security incidents of social media occur frequently.To regulate the protection of personal information of social media from the legal policy text,and protect the legitimate rights and interests of personal information and the security interests of public,our paper reviews the personal information protection policy of social media.[Method/Process]This paper took 16 social media with the largest number of users and the highest frequency of use as the research object.Through content analysis,the paper analyzed the personal information protection policy of social media from the perspectives of the collection and use of personal information,reminders of cookies and related technologies,personal information storage and protection,transfer and disclosure of personal information sharing,personal information processing rights,and protection of underage personal information.[Result/Conclusion]At the government level,it is urgent to strengthen the implementation of the "Information Security Technology and Personal Information Security Code" and other laws and regulations.At the enterprise level,it is necessary to carry out internal institutionalization of personal information protection,combine the service and product features to refine the content of personal information protection policies,and adopt the real name system to improve underage personal information protection.

[Purpose/Significance]Evaluation science of China's discourse power is an important task in the study of philosophy and social science with Chinese characteristics.It meets the realistic needs of China's system,road,theory,S&T development,management and decision-making,evaluation theory and practice and coordinated development of society in the new era.[Method/Process]Using bibliometrics,content analysis and network survey,this paper systematically summarized the relevant research literature on the evaluation science of China's discourse power at home and abroad.[Result/Conclusion]The article aimed to clarify the development context and current situation of the research in this field,and found out the shortcomings.Finally,this paper put forward some strategies to promote the evaluation science of discourse power in China,with a view of providing reference and suggestions for the follow-up related research.

[Purpose/Significance]With the increasing research interest in the digital humanities field year by year,the exploration of the research status of the digital humanities field in China is helpful to reflect on the lack of clear development direction,so as to provide reference for relevant research.[Method/Process]Based on the data sources of digital humanities literatures collected by CNKI from 2001 to 2018,this paper used bibliometrics to conduct visualized analysis and content mining on the time and space,cited literatures with high frequency and key words of digital humanities research in China,identify research frontiers,and comprehensively understood the research hotspots and development trends of digital humanities in China.[Result/Conclusion]At present,there were four research hotspots in the field of digital humanities in China;The distribution of digital humanities showed the trend of spreading from library information archives to other humanities.

[Purpose/Significance]As the key technology and basic industry for the Artificial Intelligence,Machine Learning receives unprecedented attention and rapid development.Thorough research on its technology trend and competitive posture will provide a scientific basis for the decision making of both relevant industry policies and business strategies.[Methods/Processes]Based on Derwent innovations index and Web of science database,this paper used patent statistical analysis,bibliometric analysis and social network analysis methods to analyze development status of this technology field from three aspects:the technology development stage,the core and hot technology field identification and the comparative analysis of major competitive countries,in order to clarify the strengths and weaknesses of China,and explore future development opportunities.Finally,based on the comprehensive analysis,this paper proposed corresponding countermeasures and suggestions for technological research and development,market layout and industrial development in the field of machine learning in China.[Results/Conclusion]Machine learning technology was in a rapid growth period,and China was also in a period of rapid development;there were shortcomings in the layout of technology structure in China;the United States had the strongest patent activity,and China was also a technology activist;the United States had the highest patent quality,and China was quite different from it;Internet companies were an important driving forces;the hotspots included intelligent diagnosis,autopilot,education aid,speech recognition,computer vision,and so on;the core areas included sequencing,learning,knowledge processing,search,fuzzy logic systems,and expert systems,and so on.