A Comparative Analysis of the Current Status and Trends of Domestic and International Privacy Protection Research—CiteSpace-Based Bibliometric Study (1976-2022) ()
1. Introduction
Digital marketplaces have been understood as social networks, large marketplaces, and any digital platform that identifies individual user traffic in the form of online communities (Liu & Lai, 2022; Oberg & Alexander, 2019). If these datasets are studied using artificial intelligence, machine learning, deep learning, or big data (BDA) analysis (Kar & Dwivedi, 2020), user behavior can be better predicted, but also increases the risk of user privacy violations in the digital ecosystem (Gutierrez et al., 2019). Artificial intelligence fundamentally changes the way human production and life, and a freer, more open and transparent space is revealed (Xu, Xiang, & He, 2021). With leaps and bounds in information technology, the epidemic has enabled people to work, socialize, and shop online, but inadvertently increased privacy concerns and personal privacy protection behaviors (Sundaram & Shetty, 2022). Privacy has never been more important or eroded than now, and information privacy security has become the most important ethical, legal, social, and political issue of our time (Ozdemir, Jeff Smith, & Benamati, 2017), and privacy protection needs urgent attention and improvement.
Privacy protection has been studied by scholars using CiteSpace, and the earliest documented in China is (Zhu, 2013), revealing its development patterns and trends, limited to the extremely small amount of privacy protection literature at that time, using only foreign data for analysis. Later (Gu & Fan, 2018; (Feng, Yan, & Hu, 2020; Wang, 2021), etc. only based on domestic or (Xu, Hou, & Hu, 2020; Shu & Liu, 2021) only based on foreign data for research in the field of privacy, the research topic focuses on privacy, which is broader and lacks comparative research. Therefore, based on the unfocused content and incomplete analysis of previous studies, this study uses CiteSpace to analyze the existing literature related to privacy protection at home and abroad, reveal the dynamics of privacy protection research, and provide a thinking path for researchers to understand the current state of research and carry out in-depth research (Table 1).
CiteSpace (Citation Space), developed by the internationally renowned information visualization expert Professor Chaomei Chen, analyzes the potential knowledge contained in the vast scientific literature in the context of scientometrics, data and information visualization, with multiple, time-sharing and dynamic citation visualization. Through a clever spatial layout, the evolution of the field is centrally presented on the knowledge graph of the citation network, and the research frontiers characterized by the citation node literature and co-citation clustering on the graph as the basis of knowledge are automatically identified, and the graph is highly interpretable (Li & Chen, 2017).
![]()
Table 1. Privacy-preserving research articles based on the CiteSpace Bibliometric approach.
2. Information and Methods
2.1. Data Sources
In China, the data sources were “Core Journals”, “CSSCI” and “CSCD”, and the relevant literature from the beginning of the database to August 2, 2022 was searched with the following search formula The search formula was “Subject = privacy protection or privacy concern”, and the language was limited to “Chinese” (Figure 1). The data of Web of science core collection database science citation index (SCI), social sciences citation index (SSCI) were used abroad, and the relevant literature from the beginning of database construction to August 2, 2022 was retrieved in order to study the privacy protection issue more comprehensively. The search formula of Web of science is TI = “privacy protect*” OR TI = “privacy concern*”, the language was limited to “English”, the type of article was “Article” AND “Review Article The language is limited to “English” and the type of article is “Article” and “Review Article” (Figure 2).
![]()
Figure 1. Chinese literature keyword search result chart.
![]()
Figure 2. English literature keyword search result chart.
2.2. Adoption and Exclusion Criteria
The criteria for adoption were the relevance of privacy protection and privacy concern studies, while those with missing key information, those judged irrelevant by multiple researchers after reading and discussing, and those with conference abstracts, press releases, litigation documents, editorial materials, etc. were excluded and not adopted.
2.3. Data Processing Analysis
The plain text format (.txt) of the English and Chinese literature was exported and saved, and the researchers removed them according to the adoption and exclusion criteria, after which CiteSpace 6.1.R2 (64-bit) software was used to transform and remove duplicate literature (Figure 3, Figure 4), analyze the number
![]()
Figure 3. Chinese literature data funnel diagram.
![]()
Figure 4. English literature data funnel diagram.
of publications, keywords, national collaboration networks, journal co-citations, etc., and draw visual maps. Parameter setting index: According to the privacy protection and privacy concern themes of Chinese and English literature to build the library to start the search, the time range of English literature was set to 1976-2022, and Chinese to 1997-2022. The time period separation was set to 1 year, and the data inclusion criterion was “TOP N = 50” (the data with the top 50 occurrences in each time slice). In order to avoid the complicated content of the plots, “pruning sliced networks”, “pruning the merged network” and “pathfinder” were checked to simplify the plots, and other parameters were left as default values.
2.4. Graph and Metrics Analysis
The node size represents the content of the co-occurrence map. The node size of author collaboration mapping represents the number of papers published by authors, institutions or countries/regions, the node size of network mapping of topics, keywords and scientific categories represents the frequency of occurrence, the size of nodes in co-citation analysis mapping reflects the number of co-citations of literature, authors and journals, and the connecting line reflects the corresponding strength (Li & Chen, 2017). The centrality reflects the closeness of the connection between nodes, and the higher centrality indicates the greater role played by the node.
3. Results
3.1. Annual Publication Volume
The number of publications can measure the hotness of the research field and the progress of the research, and the comparison of the number of publications can more intuitively see the degree of attention and changes in the trend of the research hotspots of the field by researchers at home and abroad. After screening and de-weighting, 4573 and 8223 Chinese and English literature were included respectively, and the annual publication volume is shown in Figure 5, with a fit of 0.8844 and 0.974, respectively, and the closer the fit to 1, the higher the credibility of the trend line (Guo & He, 2019). The English literature was first included in 1976, 21 years before the Chinese literature. Both domestic and foreign scholars privacy protection-related literature is growing (due to the inclusion of less literature in 2022, so the number of articles issued in that year decreases), and in recent years there has been an explosive growth abroad, reflecting from the side that scholars are more enthusiastic about the research on privacy protection and privacy concerns.
According to each key point of research in the development of privacy protection, it can be divided into three stages.
1) Initial stage (1976-2004): A small number of scholars at home and abroad researched privacy protection issues in this stage, with an earlier start abroad.
![]()
Figure 5. Comparison of the annual publication volume of Chinese and English documents from 1976 to 2022.
The emergence of the Internet had a certain impact on privacy protection, but the relevant knowledge and theories have not yet formed a systematic framework and structure, and WOS and CNKI core collection databases contain less literature, and the annual statistics at home and abroad do not exceed 100 articles. (Minsky, 1976) was the first to conduct privacy protection research, the study pointed out that privacy protection was completely ignored in the database literature, and the programmers maintaining the information system had almost unrestricted access to all the information in it, so deliberate solutions were introduced as an independent dimension of privacy protection.
2) Development phase (2005-2016): this phase is in a stable growth trend at home and abroad, and foreign literature is still more than domestic literature, but the gap between them is not large. The research fields of privacy protection focus on computer science, engineering, telecommunications, and business economics, etc. There are relevant studies on user online privacy, youth privacy issues, and privacy protection models in this phase. The rapid development of information technology and the popular use of e-commerce have led companies to provide personalized services to consumers based on online data. However, the extensive collection and use of personal information has led to serious consumer concerns about privacy invasion and raised the issue of personalized privacy trade-offs (Lee, Ahn, & Bang, 2011).
3) Outbreak phase (2017-2022): The volume of privacy protection publications began to increase dramatically, with foreign countries pulling away from domestic countries by a large margin, and the volume of foreign publications is almost twice as large as domestic ones so far in 2019. 2019 saw the outbreak of the new crown epidemic, with various privacy security issues exposed and the Covid-19 pandemic increasing privacy protection concerns (Sundaram & Shetty, 2022). China has also accelerated the process of privacy protection-related legislation, continuously introduced relevant national standards, industry standards, and other related management methods, and strengthened the security enforcement inspection of APPs. From the side, it also reflects that scholars' concern for privacy protection has increased substantially in the context of the epidemic.
3.2. Author Analysis
The visual analysis of posting authors can find the core authors as well as collaborating teams in the field of privacy protection research. The circle represents the node, the higher the intermediary centrality of the node the darker the color of the outermost circle will be, the higher the intermediary centrality indicates the greater the role played in the linkage network. Connected lines represent collaborative relationships, and the color and thickness are the time and number of collaborations; the lighter the color, the more recent the time (Rong et al., 2022).
The co-occurrence mapping of authors in Chinese literature is shown in Figure 6, which has more sub-networks with more Chinese research groups, but the core teams do not form network linkages with each other and do not cooperate closely. The core team represented by Jing Yang, Bing Zhang, Jing Xie, and Jianpei Zhang, with a surge of 11 publications in 2014, proposed a privacy-preserving algorithm for personalized trajectories based on typical correlation analysis to address the problem of personalized trajectory privacy protection needs, which better respects the privacy protection wishes of data generators and also obtains higher quality trajectories (Li et al., 2015). The core team of Lei Zhang, Chunguang Ma, Bin Wang, Songtao Yang, and Changli Zhou, with a large volume of publications in FY17, studied the privacy-preserving model design established to produce an indistinguishable privacy-preserving method based on the association probability of location offset (Zhang et al., 2017). And the team formed by Yingjie Wu, Weiwei Ni, Geng Chen, and Zhihong Chong analyzed the relationship between preserving the granularity of clustering features and the usability of clustering and the security of privacy protection from the perspective of preserving the granularity of clustering features, and pointed out some difficult problems to be solved and future development directions in the field of clustering-oriented data hiding publishing based on the in-depth comparative analysis of existing technical methods (Ni et al., 2012). The top ten rankings of domestic authors’ publications are shown in Table 2, with all the centrality being 0. Our authors do not cooperate closely in this field of research. The author with the most publications is Yang Jing, who proposed a privacy-preserving method based on random projection technology to solve the dimensional disaster problem in privacy-preserving data mining (Yang, Zhao, & Zhang, 2013).
![]()
Figure 6. Author’s cooperation network of Chinese literature.
![]()
Table 2. Top10 authors of Chinese literature by publishing volume.
The co-occurrence mapping of English literature authors is shown in Figure 7, which shows that the foreign literature authors cooperate very closely and the core research teams are also closely connected with each other. According to the size of the cooperation network and the order of closeness, the volume of publications of English literature authors is nearly 6 times more than that of Chinese literature authors. From Table 3, we can see that the author with the highest total number of publications is ZHANG Y with centrality 0.02. This author’s team
![]()
Figure 7. Author’s cooperation network of English literature.
![]()
Table 3. Top10 authors of English literature by publishing volume.
researches ways to protect privacy while providing accountability, and develops the Accountable Monroe Coin model, which isolates the three roles of users, trusted registrars, and trusted regulators. Accountability enables the trusted regulator to reveal the identity of the signer, and only the trusted regulator can track the user’s public key as needed (Zhang & Xu, 2022). Notably LI Y saw a spike in the number of articles in 2021, reaching a maximum of 40 articles. The team concentrates on keyword research in privacy protection, and to avoid keyword privacy leakage in both traditional models, the team constructs new global indexes to protect keyword privacy from insider attackers (Li et al., 2021). Through visual mapping analysis, the study found that the top ten authors of English-language literature are Chinese, mostly collaborating with foreign authors on their publications. In January 2018, the Blue Book on the Development of Chinese Scientific and Technological Journals (2017) counted the inflow and outflow of SCI papers from 14 major paper-producing countries in the world, and China was one of the countries with the most serious loss problem, second only to Portugal. The number of papers published in domestic SCI journals is only 0.12 times the number of papers published by Chinese scholars, meaning that at least 88% of China’s SCI articles go to foreign journals (China Association for Science and Technology (ed.), 2018).
3.3. Analysis of Country and Institutional Cooperation Networks
3.3.1. National Cooperation Network
The results were analyzed by combining the names of the countries involved in the web of science core collection source data and their abbreviations, and the results are shown in Table 4. The top 5 countries in terms of number of publications are China (3016), the United States (2640), Australia (513), the United Kingdom (464), and South Korea (452), and the highest centrality is the United States (0.34), and the high centrality represents the node then links the size of the role of the entire network. The presence of purple color in the nodes in the map indicates that the country or institution represented has a high intermediary centrality (Figure 8).
![]()
Table 4. The top ten countries in the national cooperation network.
3.3.2. Collaboration Network of Issuing Institutions
Collaboration mapping can evaluate the academic influence of countries, institutions or researchers from a new perspective, discover the social relationships between them, and assist us in finding countries, institutions or researchers worthy of attention (Chen et al., 2015). Research from the perspective of regions and overall institutions can better evaluate the research performance and objectively assess the research strengths to complement each other. From Figure 9, it can be seen that there is a relatively close cooperation between various research institutions in foreign literature, and only one of the top ten institutions in terms of the number of articles issued is from Singapore (Table 5), while the others are from China, represented by Xidian Univ (179 articles issued) and Chinese Acad Sci (centrality 0.05), Xidian University of Electronic Science and Technology is a university with mainly information and electronics disciplines, coordinated development of multiple disciplines, and the first batch of first-class cyber security college construction demonstration projects.
The research institutions of privacy protection Chinese literature (Figure 10) also have some cooperative networks, but they are sparse compared to those of English literature, and there is a lack of cooperation among institutions with high centrality (Table 6). The research institution with the highest number of publications in Chinese literature is the School of Computer Science and Technology of Nanjing University of Posts and Telecommunications, followed by the School of Computer Science and Technology of Harbin Engineering University, the School of Information Management of Wuhan University, and the School of Information of Renmin University of China, etc. The highest centrality is the Institute of Information Engineering of the Chinese Academy of Sciences, which plays a high role as a bridge. Based on the reading of institutional literature, it is known that the research at the School of Computer Science, Nanjing University of Posts and Telecommunications focuses on the differential privacy protection mechanism in the context of the Internet of Things and blockchain privacy leakage. With the development of social networks, the security protection of personal privacy data such as users’ social relationships needs to be addressed, and the institution has proposed various differential privacy protection schemes, such as BCPA, horizontal federal PCA differential privacy data publishing algorithm, and non-interactive differential privacy model dp-noisy (Huang et al., 2019; Zhu & Yang, 2022; Huang et al., 2020).
![]()
Table 5. Top 10 publishing institutions for English literature.
![]()
Figure 9. The cooperation network of English-language document publishing organizations.
![]()
Figure 10. Cooperation network of Chinese document publishing organizations.
![]()
Table 6. Top 10 publishing institutions for Chinese literature.
3.4. Keyword Analysis
3.4.1. Keyword Co-Occurrence Analysis
The larger the graph nodes represent the keywords with more occurrences, and the frequency of keywords reflects the hot spots in the field of privacy protection research. After combining and eliminating synonyms and irrelevant words in keywords, co-occurrence and clustering analysis, and extracting the top ten keywords in the keyword co-occurrence analysis of Chinese literature (Table 7), we can see that the hot spots of privacy protection research in China are focused on differential privacy, big data, blockchain, k-anonymity, privacy, cloud computing, location privacy, etc. (Figure 11). According to the top ten keywords in frequency in the keyword co-occurrence analysis of foreign literature (Table 8), it can be seen that foreign privacy protection research hotspots are concentrated in model, Internet, security, trust, data privacy, differential privacy, algorithm, location privacy, information privacy, etc., with high centrality of Internet (0.71), user behavior (0.71), and information privacy (0.68) (Figure 12).
![]()
Table 7. High-frequency keywords and centrality of Chinese literature.
![]()
Table 8. High-frequency keywords and centrality of English literature.
![]()
Figure 11. Frequency and centrality of key words in Chinese literature privacy protection research.
![]()
Figure 12. Frequency and centrality of key words in English literature privacy protection research.
3.4.2. Keyword Clustering Analysis
After the cluster analysis of Chinese and English literature, the final English literature Q = 0.8505 > 0.3000, S = 0.9558 > 0.5000, and Chinese literature Q = 0.8424 > 0.3000, S = 0.9555 > 0.5000, indicating that the clustering effect of Chinese and English literature is convincing. Chinese and English formed 0-15 clusters, respectively, and keyword clustering analysis used the log likelihood rate (loglikehood rate. LLR) method to group the more obvious clusters and the keywords included, and the research in Chinese literature (Table 9) in recent years has been biased toward information privacy issues in the era of artificial intelligence, and focused on the regulation of data release and data sharing, and through the relevant The research in the Chinese literature has focused on the regulation of data release and data sharing, and the protection of user privacy through relevant privacy policies (Figure 13). The research in English literature (Table 10) focuses on the privacy disclosure and privacy protection of online users in social media on the mobile network side, and studies privacy protection means in all aspects through information technology, such as access control, structural model, same-station encryption, searchable encryption, etc. (Figure 14). The common research hotspots in Chinese and English literature are differential privacy, location privacy, trust, information security, privacy paradox, and social networks.
![]()
Table 9. Main cluster analysis of Chinese literature keywords.
![]()
Table 10. Main cluster analysis results of English literature keywords.
![]()
Figure 13. Clustering diagram of Chinese literature keywords based on LLR method.
![]()
Figure 14. Clustering diagram of English literature keywords based on LLR method.
3.5. Research Frontier Analysis
Changes in research themes and hotspots in a field can be studied from keyword bursts (Wang et al., 2020), and the black horizontal line markers represent the beginning and end of keyword bursts; the longer the length of the burst, the longer the keyword hotspot lasts, the stronger the research frontier, and the higher its importance and the degree of attention it receives (Han et al., 2022). CiteSpace 6.1.R2 (64-bit) software was used to analyze the English and Chinese literature for the emergent words, and 96 English and 61 Chinese emergent words were obtained. The words privacy protection, privacy concern were hidden in the search formula, and the top 23 emergent words by emergent year are shown in Figure 15 in English and Figure 16 in Chinese.
![]()
Figure 15. Top 23 emergent words in English literature.
![]()
Figure 16. Top 23 emergent words in Chinese literature.
Informed consent is the keyword with the longest duration of foreign keyword emergent analysis (1994-2016), and the highest emergent intensity as of August 2022 is information privacy (23.54), followed by e-commerce (16.73) and trust (14.94). Foreign studies in recent years have focused more on encryption, location privacy and anonymization.
In China, e-commerce is the keyword with the longest duration of domestic keyword emergence analysis (2002-2014), “artificial intelligence”, “smart contract”, “Data governance” and “blockchain” have been continuing since their emergence in 2020, and are expected to become research hotspots in the field of privacy protection in the future, which researchers should pay attention to and focus on. Although domestic research on privacy protection and privacy concerns started later than foreign research, and there is a certain lag in research, the growth trend of the number of articles published in recent years is synchronized with foreign research, which indicates that domestic scholars’ enthusiasm for research in this field is increasing.
4. Discussion
4.1. Comparison of the Current Status of Domestic and International Research
After the outbreak, privacy protection and privacy concerns have exploded in international attention, and have remained above 1000 articles per year since 2019. The authors ranking 1st in English publications and the top 6 institutions are all from China, indicating the importance of Chinese research in this field, but it cannot be ignored that the centrality of the United States and the United Kingdom is very high, while that of China is only 0.05, and the level of research needs to be greatly improved. Compared with foreign countries, the number of publications in China has also been growing in recent years, with a more moderate growth trend and significantly lower than foreign countries. A visual mapping analysis of the number of publications and institutional cooperation networks by country shows that the research in the field of privacy protection and privacy concerns is in a key position in China, the United States, Australia, the United Kingdom, and South Korea.
The analysis of the co-occurrence mapping of authors and collaborating institutions shows that foreign institutions and authors have built a close collaborative network and the number of publications has increased substantially and rapidly. In contrast, the collaboration among research institutions and scholars in China is not close, and there are few large professional research teams, which makes it difficult to conduct large-scale and high-quality research in China. Moreover, a large number of Chinese scholars cooperate closely with foreign scholars and publish through foreign channels. In this regard, the author suggests that more domestic and foreign academic exchanges should be carried out, actively learn from foreign experience, and publish strong national cooperation strategies. Research institutions at the core of our country should take the lead and take the initiative to actively develop cooperation plans with neighboring research institutions and gradually expand the scale of research groups.
4.2. Domestic and Foreign Research Content Partially Overlaps and Has Its Own Focus
Based on the analysis of keyword co-occurrence, high-frequency keywords, keyword clustering and visual graphs of emergent keywords at home and abroad, foreign countries focus on patterns, security, trust, data privacy, personalized privacy, etc., and begin to study privacy protection earlier, and have profound theoretical support, privacy protection systems and models for privacy protection. Chinese scholars’ research hotspots are focused on differential privacy, big data, blockchain, location privacy, etc. Compared with foreign countries, the focus is on computer science for privacy protection. The comparison of emergent words can see the different perspectives of privacy protection at home and abroad. Foreign countries focus more on the subjective feelings of online users’ privacy, such as informed consent, trust, attitude, etc., and control privacy infringement and other problems by establishing privacy protection models. Domestic is more concerned with data governance, data computing, machine learning, etc. After visualization and analysis we are able to see that domestic is starting to focus on the privacy of AI online users.
4.3. Research Features and Limitations
This research uses the current scientific and cutting-edge visualization mapping software Citespace 6.1.R2 (64-bit) to conduct the research. The mapping is a clear and intuitive way to understand the history, development and trends in this field at home and abroad, and has a high reference value. The use of domestic and foreign core data for comparative research ensures the comprehensiveness of the study. Due to the large volume of literature, only the core data of China Knowledge Network was used for Chinese literature, and the database source is relatively single. With the increase of research on privacy protection and privacy concerns, the database of Chinese literature will be enriched at a later stage and further improved for analysis in the future.
5. Conclusion
The digital age has increased the speed of the network, a multi-dimensional sensory world has arrived, and the efficiency of transmission has increased ten thousand times. Companies and governments around the world are increasingly relying on technology to collect, analyze and store personal data. If location information and trajectory data are directly released without any control, and analyzed and mined at will, it will lead to serious social problems due to the leakage of a large amount of personal privacy information. It is clear that (Sundaram & Shetty, 2022) the technological changes caused during Covid-19 affect privacy concerns and protective behaviors, privacy efficacy increases privacy concerns and protective behaviors, and online video chat software users’ privacy concerns and protective behaviors increase accordingly. Through the analysis of domestic and international data (Wan et al., 2016), location privacy protection technology gradually becomes a research hotspot and a large number of research results emerge, location data can expose a large amount of personal information, social relationships, religious beliefs, behavioral habits, health status, political tendencies, etc., so we should pay enough attention to it. With the leapfrog development of social networks, a large number of privacy leaks occur frequently at home and abroad, in addition to the technical aspects to solve the privacy protection problem, it should also be solved from the integration of various aspects such as enterprise management, consumer psychology, value perception, and privacy paradox.