The method used in this paper is a co-word analysis. Co-word analysis technique, based on the calculation of the rate of all co-occurrence of words, provides the possibility of visualizing the structure of subjects and concepts (24). This method is suitable for detecting communications, tracking structural changes, and developing the conceptual, social network, and essential links (25). Therefore, this method was used to discover the structure of global knowledge, hidden patterns, and emerging Coronavirus events.
The statistical population is all English language articles published about Coronavirus from 1970 to 2019. Using this interval can comprehensively show the intellectual structure of Coronavirus. The Web of Science Core Collection (WOSCC) was utilized to access the research community. It is the world's most prestigious, widely used, and the oldest citation database (11). Since 1900, with full cited reference indexing and, in comparison with other databases, Web of Science has provided more valuable and comprehensive documents for downloading in various scientific disciplines (26). The MeSH (Medical Subject Headings) browser is used to identify the Coronavirus main keywords. After consulting with experts, the search strategy was designed and developed in the Web of Science advanced search. The search strategy is as follows.

"Coronavirus" OR "ALPHACORONAVIRUS" OR "BETACORONAVIRUS" OR "DELTACORONAVIRUS" OR "GAMMACORONAVIRUS" OR "Bulbul coronavirus HKU11" OR "Coronavirus HKU15" OR "Munia coronavirus HKU13" OR "Rabbit Coronavirus" OR "Thrush coronavirus HKU12".

The advanced search results using the Title Tag (TI) were limited to ARTICLES, MEETING ABSTRACT, PROCEEDINGS PAPER, and REVIEW in terms of the document type. In terms of language, retrieved documents were limited to the English language for the period 1970-2019. The reason for choosing the period 1970-2019 was the lack of any publications about Coronavirus prior to 1970. A total of 4102 Keywords extracted from indexed documents were retrieved through advanced search. The keywords were studied in a process that included editing, integrating, deleting, and mending. After reviewing the process, the number of keywords decreased to 4058. Ravar Matrix: Ravar PreMap was used to refine the data and create a co-occurrence matrix.
To prepare the matrix for co-word analysis, researchers selected the threshold. After performing some tests, Bradford's law was utilized in the final analysis. In previous co-word analysis studies, different thresholds were used to include the top keywords in the final analysis (13, 27). The threshold of this article determined using the Bradford Law on 11; in other words, keywords that have been repeated more than ten times. Using the threshold, 86 high-frequency keywords were identified and analyzed in the co-word analysis (Table 1). To perform the co-word analysis, hierarchical clustering was done with the Ward method and Squared Euclidean distance calculation using SPSS 16 (SPSS Inc., Chicago, Ill., USA). Then, the dendrogram diagram was drawn with 19 subject clusters (Figure 2 and Table 2). The use of the Ward method for hierarchical clustering analysis has been performed in many previous co-word analysis' studies (28-32).
Subsequently, the co-word matrix characteristics, such as Centrality and Density, were analyzed using the UCINET software to obtain more information about the Coronavirus's co-word structure. If the node has many relationships with other nodes in a network, it has a higher centrality and an important place. Density means more cohesion and more internal interconnection between nodes. Density is a research area that demonstrates its ability to sustain and develop. The Centrality and Density of each of the subject clusters can be demonstrated in the form of a Strategic Diagram, revealing the Coronavirus's state and evolutionary trends. The Strategic Diagram's horizontal axis was Degree / Centrality, and its vertical axis was density (27, 32). In Strategic Diagram, the higher the centrality of a cluster is, the more critical that cluster is in the research topic, and the higher the density of a thematic cluster is, the more capable or mature that cluster is. Therefore, in a strategic chart, the X-axis represents the centrality, and the Y-axis represents the Density (Figure 3). The Strategic Diagram can be illustrated in 4 quadrants, each with different densities and centralities, and each cluster in each quadrant has a different status. Finally, a visualization was performed using VOSviewer1.6.11.
 

Coronavirus Top keywords based on co-word analysis
According to Table 1, the top 20 most frequently keywords in the Coronavirus subject area were presented based on the co-word analysis. The "Severe Acute Respiratory Syndrome (SARS)," with 276 frequency, had the highest frequency among all keywords. MERS-COV and SARS-COV were also ranked second and third with a frequency of 265 and 259, respectively. Table 1 showed the importance of these keywords for researchers at this time.

Table 1. Ranking the top 20 keywords in the Coronavirus  (1970-2019)

Figure 1. Coronavirus High-frequency Keywords network structure

Figure 1 shows the high frequency keywords in Coronavirus subject area. Hot topics were based on the high frequent Coronavirus related keywords, including 86 keywords with a frequency above 10. These subjects were shown in the form of a network with VOSviewer in Figure 1. The size of the nodes indicates the weight of the scientific publications of each subject. The colors represent the formed nodes. In the map drawn with VOSviewer, the topics were networked according to existing algorithms. The colors indicate the subjects, and the diameter of the lines indicates the connection between the nodes. Hot subjects with higher co-occurrence are centered on the network, and topics with lower co-occurrence are around the network.
 
Coronavirus hierarchical clustering in the world
Cluster analysis of scientific publications in the last half-century has led to 19 clusters in Coronavirus. Table 2 also shows the Coronavirus clusters along with the keywords for each cluster.
 

Table 2. Coronavirus subject clusters based on clustering analysis

Figure 2. Coronavirus subject clustering dendrogram in the World (1970-2019)

The First cluster has two keywords. The tenth high frequent keyword is "Feline Coronavirus (FCOV)," and "Feline infectious peritonitis (FIP)" is the twentieth high frequent keyword. They are located in the first cluster. Researchers have noted the importance of Coronavirus and the Peritonitis of infection. Transmissible Gastroenteritis Virus (TGEV) is the most frequent keyword in the second cluster. The keywords in the 2^nd cluster illustrate the issue of Gastroenteritis and the transmission of crowned Coronaviruse from animals; animals such as Porcine and Bovine. The third cluster has five keywords. Bovine Coronavirus (BCOV) is the sixth most frequent keyword in this research. According to the third cluster subjects' identification and analysis, "Turkey Bovine Coronavirus Immunohistochemistry" is the main subject of this cluster. The fourth cluster has the highest number of keywords among all clusters. It consists of nine keywords. Keywords for this cluster are "Enzyme-linked Immunosorbent Assay (ELISA)", "Real-time RT-PCR", and "Canine coronavirus (CCoV)". Considering the presence of Polymerase Chain Reaction keywords and ELISA, Serology, and Real-time RT-PCR tests, it can say that the main topic of this cluster is "Coronavirus Laboratory Testing". The fifth cluster has two keywords. The seventh most frequent keyword (Infectious bronchitis virus" (IBV)) is in this cluster. The main subject of this cluster is the “Avian Coronavirus infectious bronchitis virus.” The sixth cluster has three keywords. “Porcine Delta Coronavirus (PDCOV)” and “Phylogenetic analysis” are the eighth and the 11^th most frequent keyword in this cluster, respectively. The main subject of the 6^th cluster is the “prevalent phylogenetic analysis of porcine Coronavirus”. Five keywords involved in the formation of the seventh cluster. Analysis of the subjects of this cluster showed that the main subject of this cluster is “Molecular Epidemiology of Recombination Human Coronavirus”. The eighth cluster has seven keywords. Due to the presence of “Epidemiology” and “Human Coronavirus” in this cluster, the main subject is “Human Coronavirus and viral infections”. The ninth cluster includes three keywords, Interferon and Antiviral, Ribavirin. So, the subject of this cluster is “Coronavirus Drug Therapy”. Three keywords have been involved in the formation of the tenth cluster. “Spike Protein” is the fifth most frequent keyword with the rank of sixty-eight among all keywords in this research. The keywords in the tenth cluster indicate that “Coronavirus antibodies” have been the subject of studies of researchers. The eleventh cluster has five keywords. The second and the fourth most frequently keywords in this domain are “MERS-COV” and “Middle East Respiratory Syndrome”. Based on these keywords, this cluster is about the “Outbreaks of zoonotic diseases caused by viruses MERS-COV”. The 12^th cluster consists of four keywords. Coronavirus infection causes demyelination of the Central Nervous System (CNS). The keywords indicate that the subject of the 12^th cluster could be “Coronavirus infects in the CNS”. This cluster has three keywords. Analyzing the subjects of the keywords of the 13^th cluster suggests that “Structures and Mechanisms of Coronavirus Cell” well illustrates the central theme of this cluster. “Envelope Protein” and “Ion channels” are the keywords of the fourteenth cluster. CoV envelope (E) protein is involved in the life cycle of the virus. According to the keywords in this cluster, the “CoV life cycle” indicated the subject of the 14^th cluster.
The cluster fifteen, like cluster four, has nine keywords. The third most frequent keyword is "Severe Acute Respiratory Syndrome Coronavirus (SARS-COV)" with a frequency of 259 in this cluster. The key theme of this cluster is related to "Pathogenesis and SARS-COV". This cluster has seven keywords, and keywords such as Pathogenesis, Pathogenicity, Viral replication, and infection present in the 16^th cluster. By analyzing keywords, the main subject is "Pathogenesis of Viral Infections". The 17^th cluster, consist of six keywords. This study's most frequent keyword is "Severe Acute Respiratory Syndrome (SARS)", with a frequency of 276 in this cluster. Vaccine, the thirteenth high frequent keyword, is also in the cluster. "SARS and Vaccination" represent the central concept of the seventeenth cluster. The two keywords are "Monoclonal Antibody and Epitopes" in the 18^th cluster. So "Immunology and Coronavirus" is the main subject of this cluster. The 19^th cluster has six keywords, and Nucleocapsid protein is the ninth most frequent keyword in this cluster. M-CoV is a type of Coronavirus that comes with the names of proteins. Hence the content of this cluster deals with the "structure of Coronavirus and Epidemic murine illness."


Coronavirus subject clusters Density and Degree/Centrality
The data in Table 3 showed the density and centrality of the nineteen Coronavirus subject clusters. Using UCINET software, the Degree / Centrality and Density indices of each cluster were calculated. Also, the average of these two indicators were calculated for all clusters. Clusters 11, 15, and 17 had the highest Degree / Centrality, respectively, and Categories 1, 5, and 11 had the higher densities than other clusters (Table 3).

Table 3. Coronavirus subject clusters density and Degree/ Centrality (1970-2019)​​​

Coronavirus Strategic Diagram
The Strategic Diagram's origin was determined by the Degree / Centrality and Density averages of the clusters. Degree / Centrality determined the interaction power of each cluster on the X-axis. The higher the centrality of a cluster is, the more critical and central it is. The internal relationship of the clusters with density in the Y-axis were also shown. The higher the density of a cluster is, the greater the cluster's ability is to maintain and develop (13, 33, 34). A cluster located within a strategic diagram makes it characterize in the context of the whole discipline. In quadrant I; clusters were both coherent and centric to the field. These mainstream clusters figured the focus of a large section of the network. The quadrant II clusters were also coherent; however, they are likely to be specialized and separated from the overall focus. The clusters in quadrant III were in infusion, showing the emerging or declining sections of the network. Finally, quadrant IV contained the clusters that illustrate a common and broad focus or have not yet completed but have the potential to be considered as a primary network focus (32, 35). The data in Figure 3 shows that in quadrant I, cluster 11; in quadrant II, clusters 1, 2, 5, and 6; in quadrant III, clusters 12, 8, 3, 18, 7, 9, 19, 4, and 16; and in quadrant IV, clusters 15, 17, 10 and 13 are located.
 

Figure 3. Coronavirus Strategic Diagram (1970-2019)

 

Discussion

In this paper, using co-word analysis, social network analysis tools, and scientific visualization software, global knowledge structure, hidden patterns, and emerging events of Coronavirus in the last 50 years (1970 to 2019) were presented.
The findings showed that "Severe Acute Respiratory Syndrome (SARS)" has the highest frequency among Coronavirus keywords; besides, "Severe Acute Respiratory Syndrome CORONAVIRUS (SARS-COV)" is the third most frequent keyword. Therefore, SARS appears to have accounted for a large share of Coronavirus research before 2020. From November 1^st, 2002, to June 11^th, 2003, SARS-CoV spread to thirty countries and infected 8098 people, and 916 ones lost their lives as a result of SARS-CoV. Its pandemic period lasted until 2004, so between 2002 and 2004, various aspects of the disease came to medical researchers' attention, and a large number of scientific publications were published on the importance of the SARS-CoV Outbreak (36).
"Middle East Respiratory Syndrome Coronavirus (MERS-COV)" is the second most frequent keyword; the newer Coronavirus, MERS-CoV, spread to Saudi Arabia in September 2012 (37). Outbreaks in twenty-seven countries, infectious more than 2428 people, and the death of 838 were the effects of MERS-CoV. After Saudi Arabia, the disease recurred in the Republic of Korea in 2015 (36, 38, 39). In order to identify, prevent, and treat MERS-CoV, medical scientists have conducted various studies that have led to the publication of significant scientific documents.
Examination of the results showed that the increase in Coronavirus's global scientific publications follows the outbreak of the new species of the virus, the pandemic stage, and the resulting death. So, with the onset of the massive COVID-19 epidemic, the latest form of Coronaviridae, and its imminent threat to global public health (40), we will see a dramatic increase in COVID-19 global science production soon.
In this paper, in addition to co-word analysis, the hierarchical cluster analysis method was also performed. The cluster analysis results led to the formation of 19 thematic Coronavirus clusters and identified the intellectual structure of Coronavirus research. The results of this article indicated that clusters such as "Coronavirus and the Peritonitis of infection," "Gastroenteritis and the transmission of crowned Coronavirus", "Turkey Bovine Coronavirus Immunohistochemistry", "Avian Coronavirus infectious bronchitis virus", "Human Coronavirus and viral infections", "Coronavirus Drug Therapy", "Coronavirus antibodies", "Outbreaks of zoonotic diseases caused by viruses MERS-COV", "Coronavirus infects in the CNS", "CoV life cycle", "Pathogenesis and SARS-COV", "Immunology and Coronavirus", and "structure of Coronavirus and Epidemic murine illness" have been considered by medical researchers.
Tracking sciences illustrated Coronavirus research events' dynamics and the expression of hidden research patterns by identifying the structure, concepts, and components of Coronavirus outputs using the phonological analysis method. The results could be of great help to scientists and professionals who are currently researching COVID-19 to target research, given their research priorities.
The strategic diagram results in the present paper showed that the 11^th cluster "Outbreaks of zoonotic diseases caused by viruses MERS-COV" cluster has a high centrality and density. The 11^th cluster has a crucial role and is coherent and central. The topics "Coronavirus and the Peritonitis of infection", "Gastroenteritis and the transmission of crowned Coronavirus", "Avian Coronavirus infectious bronchitis virus," and "Human Coronavirus and viral infections," are also coherent but separate and do not play a central role. Subject clusters such as "Turkey Bovine Coronavirus Immuno-histochemistry", "Coronavirus Laboratory Testing", "Molecular Epidemiology of Recombination Human Coronavirus", "Human Coronavirus and viral infections", "Coronavirus Drug Therapy", and "Pathogenesis of Viral Infections" have less centrality and density. They are of little importance in the field of coronavirus knowledge discovery. Given the prevalence of COVID-19 in the Coronaviridae family, these clusters represent emerging or declining portions of the network. It is essential to consider such concepts. Understanding the structure and characteristics of COVID-19 will allow specialist physicians to discover new vaccines, medi-cations, and novel treatments. The thematic clusters "Coronavirus antibodies", "Structures and Mechanisms of Coronavirus Cell", "Pathogenesis and SARS-COV", and "SARS and Vaccination" have high centrality and low density. The clusters have not yet matured but can be a primary network focus (32).
The present study complements the novel Coronavirus (2019-nCoV) (18). Danesh and Ghavidel's research used scientometric indicators to represent the quantitative and qualitative content of coronavirus scientific publications. The results of both studies showed that the amount of coronavirus outbreak in different periods has led to the growth of scientific publications in the Coronavirus subject domain. Also, this paper's results are in line with the studies of Bonilla-Aldana et al. and (18) and Zyoud (21). In analyzing and identifying the connections between the concepts of a thematic domain in medical sciences, the present study is consistent with the Human Papilloma Virus (19). In that study, the concepts were analyzed in the form of thematic clusters, and the maturity and coherence of the topics were presented using the strategic diagram (19). According to research on the subject matter of viruses (21-24, 41), various and varied scientometrics methods have been used to analyze each topic. Important and strategic results obtained from scientometric studies, especially from the present research, make researchers and specialists in respiratory and infectious diseases familiar with the coronavirus's knowledge structure and thematic contexts. Research policy-makers can also use the results of the paper obtained in macro-planning.
The authors of the paper believe that it is necessary to conduct co-citation and co-authorship research to identify the citation network, introduce proliferated scientists, and discover global scholarly collaboration patterns. Since the outbreak of COVID-19 numerous articles were published; we suggest that all the COVID-19' s publications in 2020 be analyzed in an independent research using scientometric indicators. Undoubtedly, the results of such studies will be of interest to international medical researchers and policymakers.
 
 

Conclusion

Discovering the structure of global knowledge and its impact on the research of the Coronavirus will be able to establish a link between human knowledge and the activities of scientists and specialists during the COVID-19 pandemic. Representing and visualizing the scientific structure of Coronavirus, using hierarchy clustering, showed the breadth of concepts and the discovery of the hidden connections and patterns of Coronavirus. The maturity and cohesion of each of the Coronavirus clusters will guide Coronavirus researchers toward emerging events and pave the way for planners and policy-makers in international medical organizations.

Ethical considerations

Ethical issues (Including plagiarism, informed consent, misconduct, data fabrication and/or falsification, double publication and/or submission, redundancy, etc.) have been completely observed by the authors.
 

Acknowledgements

The authors of this article would like to appreciate the specialists in respiratory and infectious diseases for expressing valuable views and comments.
 

Conflicts of Interest

Authors declared no conflict of interests.