Objective: The purpose of this study was to analyze the topics of theses and dissertations in the Iran Scientific Information Database (GANJ) and determine criteria for selecting topics based on information organization standards. Additionally, the study aimed to identify the main challenges in selecting relevant topics.
Methodology: The research method was qualitative and its approach was comparative. Data collection involved conducting interviews and extracting relevant data from postgraduate theses and dissertations available in the Iran Scientific Information Database (GANJ). The statistical population of the database consisted of 1,112,531 documents. Sampling was done according to the Morgan table, and 1,500 documents were selected for review. The validity of the research tool was assessed using the face validity method with the input of 5 academic experts in information science. Additionally, the reliability of the assessment tool, a checklist to determine the thematic connection between the structure of theses and dissertations and the allocation of related keywords, was evaluated using the Cronbach's alpha statistical method, with a value of 0.872. The study was conducted on this sample in proportion to the size of the field and academic orientations. The analysis focused on the relevance between thematic keywords and the core concept of theses and dissertations, taking into account existing information sources to evaluate the new structure.
Findings: The findings revealed significant challenges in organizing topics, including mismatches between "topic areas" and "topic levels", absence of some "main topics", non-standardization of topics, inability to "add a topic", lack of clarity in the concept of the topic area, and insufficient "supplementary keywords". The findings show that subject keywords in a subject classification system play a crucial role in efficiently searching for information and retrieving information sources. Users typically try to choose the most relevant keyword when searching for information. If they do not find the desired results, they refine their keywords based on feedback and results to further explore by adjusting their search queries. Subject ranking is also important, as determining high-level and low-level relationships in the indexing structure can help identify relevant keywords for accurate retrieval. Using specialized dictionaries can enhance the subject area, However, this is currently not feasible.
Conclusion: The study concluded that using pre-determined keywords outside the thesaurus structure complicates subject assignment for indexers. Utilizing citation networks, techniques such as clustering in subject structure, understanding the relationship between general and specific terms, dependent terms, and the reference system can improve keyword assignment accuracy. Implementing automated information organization methods can enhance information analysis and processing, improve keyword assignment accuracy, and optimize information retrieval. The results suggest that the similarity between keywords and visualization descriptors is vital, often demonstrated through the use of specialized tools such as thesauruses. Additionally, utilizing the Web of Science classification system can significantly improve retrieval results. However, this may not be adequate for exploring smaller trends that require detailed examination. Considering other classification systems such as Scopus can be beneficial. Incorporating synonym keywords can improve the selection of subject keywords. Folksonomy systems can also enhance the subject structure effectively.