2.2. Data Extraction and Analysis

Bibliographic items cooccurrence matrix builder (BICOMB) is a basic software tool that was developed by Lei at China Medical University for text mining [15]. In this study, BICOMB was used to filter data and construct matrices. When BICOMB determines keyword frequency, it can be executed for data cleaning by grouping different expressions with the same meaning into one keyword: dental implant and dental implants and Dental implant and Dental Implant with different singular and plural forms, capitalization, and abbreviation and with or without spaces. High-frequency keywords were defined as appearing more than 60 times. Subsequently, high-frequency keywords, a dental implant binary matrix, and cooccurrence matrix of keywords were created. Next, we performed cooccurrence analysis and double-clustering analysis on the filtered data and matrix. Cooccurrence analysis uses the cooccurrence of vocabulary and noun phrases in the literature set to determine the relationship between the keywords in the field. It is generally believed that the more times a vocabulary pair appears in the same document, the thicker the connection is between the two keywords. Ucinet 6 software, which was designed by Stephen Borgatti and colleagues from the University of California Irvine, allows performing a cooccurrence analysis for social network analysis (SNA) [16]. Double-clustering analysis cluster data in both the row and the column directions of the matrix. In the resulting cluster, the data are related to each other in both the rows and the columns. We used gCLUTO software to generate mountain peak map and then to summarize and analyze the relationship between each category of dental implants. SCIMAT software was used for strategic coordinate diagram construction [17]. Further, we determined the subcategories of dental implants.