This study proposes an automated topic labeling method based on large language models (LLMs), capable of generating meaningful term and summary labels for each topic within a topic model. Two generation strategies are introduced: Concise Descriptor Labeling (CDL) and Context-Enhanced Labeling (CEL). In addition to qualitative observation and conventional measures of stability and topical relevance, this study further evaluates coverage and discriminativeness through a topic assignment task. The experimental results show that CDL tends to produce concise and general disciplinary terms, characterized by high stability and close semantic alignment with the topic descriptors. In contrast, CEL often generates specialized technical terminology; although lexical variations may occur across multiple generations, semantic consistency remains high. Though CEL demonstrates slightly better performance in terms of coverage and discriminativeness, both strategies are acceptable and complementary, allowing researchers to select the appropriate approach depending on the interpretive and application context of the topic model. Future research may focus on refining these generation strategies and exploring their integration to enhance the applicability and interpretability of topic models.
Libraries bear the responsibility of ensuring equitable access to information. With the advancement of an aging society and the implementation of the Convention on the Rights of Persons with Disabilities (CRPD), how public libraries realize “barrier-free access” and “active inclusion” has become a critical issue. This study employed a questionnaire survey targeting public libraries at all administrative levels in Taiwan to investigate their human resources, collections, reader services, facilities, and challenges in providing services for persons with disabilities. The findings revealed that Taiwan’s public libraries rely heavily on volunteers, with insufficient numbers of full-time staff, and only a few libraries have stable funding sources. Existing resources mainly consist of audiobooks, large-print books, and dual-vision books, and the collection growth remains slow. Reader services and outreach programs are diverse, with municipal libraries being the most proactive in promotional activities. However, discrepancies of facilities were found between those that technically meet accessibility standards but are not truly user-friendly, and only a small proportion of libraries have designated reading areas for people with disabilities. There is still a need to establish systematic policies, staff training, and operational guidelines to help overcome structural constraints and achieve the goal of reading equality of libraries at all levels. (Article content in Chinese with English extended abstract)
In response to the rapid changes in the global labor market and the widening gap between cacademia and industry, higher education must strengthen students’ core competencies for career development. The field of Library and Information Science (LIS) has been greatly influenced by technological advancement and the transformation of knowledge services. Cultivating LIS students with employability, career self-efficacy, and career adaptability has become an essential goal. This study surveyed students from several LIS departments in Taiwan through a questionnaire to examine the overall performance and differences among three constructs. The results indicate that students demonstrated medium-to-high levels in all three aspects. Significant differences were found in gender and grade level, while institutional level and interdisciplinary learning experience showed no significant variation. Learning initiative and self-awareness were identified as key factors influencing career development. The study recommends that LIS departments develop tiered counseling and interdisciplinary curriculum support mechanisms based on students’ characteristics and learning stages, to promote practical readiness and industry alignment. Students are also encouraged to enhance their employability, career self-efficacy, and career adaptability to strengthen professional growth and career competitiveness. (Article content in Chinese with English extended abstract)
The trend toward providing scholarly communication services stems from university libraries in Europe and the United States. However, there has been limited discussion of scholarly communication services within domestic university libraries. This study aimed to explore academic librarians’ understanding of the importance and current status of scholarly communication services, as well as their knowledge and needs for scholarly communication knowledge. This study employed a quantitative survey method, focusing on librarians from research universities in Taiwan, with a total of 122 responses. The findings indicated that several items, such as AI tools in research support and ethical issues in scholarly publishing, showed significant differences in the importance and current status of various scholarly communication services, as well as in librarians’ knowledge and professional needs. Significant differences also emerged between supervisors and non-supervisors, those with different scholarly communication responsibilities, and the size of their libraries. These findings may help research university libraries refine their scholarly communication service strategies. (Article content in Chinese with English extended abstract)
The citations to open access preprint and working paper versions of papers that are also formally published in journals are not regularly counted in global multidisciplinary proprietary citation index databases, nor are they considered in most bibliometric studies. They are, however, integrated into paper citation counts in other citation databases. It can be argued that open access preprint citations reflect a relevant part of scientific impact of a work as proxied by citation counts, particularly in scientific disciplines that make extensive use of preprints/working papers for rapid open communication. Here we present a large-scale study of these usually uncounted open access preprint citations for the Web of Science database to provide an exploratory overview of the magnitude of this phenomenon and its distribution across disciplines. To this end, we introduce a methodology to link cited reference strings referring to preprints/working papers to their matching published journal versions. We find that at least 390,000 WoS-indexed publications have 1.1 million preprint version citations which are not included in ordinary citation counts.
Conventional approaches to teaching and learning English can lead to pressure and create obstacles for learners. Using pictorial storybooks for English learning might offer students motivation to learn English. This study aimed to study several issues related to college students’ English learning, including reactions to barriers to English learning, reactions to pictorial storybooks for supporting English learning, essential characteristics of pictorial storybooks, expectations of library acquisitions and promotion of pictorial storybooks, and the design of pictorial storybooks. The subjects of the study were undergraduate students at Fu Jen Catholic University. A questionnaire with a 5-point Likert scale was employed to assess students’ responses. According to the 253 data gathered, the majority of students spent less than 1 hour per week on English learning, and they preferred visual and kinesthetic learning. The English learning barrier with the highest rating was limited lexical resources. Results reflect several points, namely the importance of integrating English pictorial storybooks to ensure enjoyment of knowledge acquisition, applicable length and amount of content for learning, expectations of multicultural topics, and appropriate design of English pictorial storybooks for supporting reading. Several suggestions are provided for fostering learning enjoyment and fulfilling students’ needs in visual and kinesthetic learning. Library development and the promotion of pictorial storybooks regarding multicultural topics are also suggested.
As the internet becomes essential to daily life, fiction writing and dissemination have also shifted into the digital realm. This study focuses on web fiction writer with experience in long-form writing, active reader interaction, and familiarity with web-based platforms. Using qualitative interviews and observations, it explores writers’ writing processes and influencing factors. Findings indicate that writing process consists of three main stages: idea incubation, plot development, and manuscript writing, accompanied by data collection, revision, and reader engagement. Influencing factors fall into two categories: personal and external. Personal factors—such as personality, emotional state, academic background, and interests—act as the core driving force behind creation. External factors—such as reader feedback, platform features, market trends, and social issues—mainly affect the direction and content during later stages of writing. Overall, web fiction writing is shaped by the interplay of multiple factors, producing a unique and dynamic creative process. (Article content in Chinese with English extended abstract)
Large language models (LLMs) have in recent years spurred research across various sectors, owing to their remarkable zero-shot or few-shot performance. This capability has become indispensable for individuals seeking to integrate these language models into their workflows effectively. In this paper, based on in-depth linguistic analyses, we explore the application of an LLM, specifically GPT-4, in generating Chinese language textbooks tailored for grade school students. This encompasses the creation of main lesson texts alongside accompanying Chinese character exercises. Experimental results suggest that the LLM-generated textbook lessons are a viable research direction. The initial outcomes demonstrate the ability of LLM to generate texts of satisfactory quality appropriate for a specified grade level. The contributions of this work include pioneering the quantitative analysis of Chinese language textbooks for native speakers in Taiwan and leveraging an LLM to automatically generate textbook content and accompanying Chinese character exercises targeted at native Chinese speakers, which is a novel approach facilitated by the development of prompts tailored to different language learning levels. The study also conducts quantitative and qualitative comparisons between machine-generated lessons and those developed by educational professionals in Taiwan.
This study adopts a mixed-methods approach to systematically examine 127 citizen science projects in Taiwan through content analysis. The analysis identifies each project's core characteristics, data collection methods, and data types, and evaluates their levels of data openness based on the FAIR Principles (Findability, Accessibility, Interoperability, and Reusability). In addition, the Taiwan Roadkill Observation Network (TaiRON) is used as a representative case, with qualitative interviews conducted to explore its data quality management strategies and the gradual transition process undertaken in response to evolving challenges. The findings reveal that most citizen science projects in Taiwan demonstrate long-term continuity and strong connections to policymaking. Data collection is primarily driven by web applications and social media platforms, fostering diverse mechanisms for maintaining data quality. In the TaiRON case, the project began by encouraging public reporting through social media to broaden participation, and progressively evolved into a more systematic and standardized survey model as data volume and community maturity increased. Furthermore, this study established open metadata for Taiwan's citizen science projects on Wikidata to enhance their discoverability and reusability. Overall, this research provides a systematic overview of Taiwan's citizen science landscape and proposes management strategies that balance data quality with openness, thereby supporting research teams and policy agencies in strengthening the contributions of citizen science to sustainable development.
Access to sufficient, high-quality data is essential for effectively training and validating machine learning classifiers. This study investigates class mapping as a data fusion strategy to enhance training data for research classification. Two versions of the Australian and New Zealand Standard Research Classification, ANZSRC 2008 FoR and ANZSRC 2020 FoR, are used to organize 179,431 documents from eight institutional repositories into plain and mapped datasets. Each dataset is divided into subsets corresponding to the division, group, and field levels of the classification schemes. Results show that 49% to 63% of documents are successfully mapped between schemes. Classifiers by Support Vector Machines (SVM), SciBERT, ModernBERT-base, and ModernBERT-large are trained to assess the effectiveness of this data fusion approach on classification performance. All models show improved performance at the three levels. ModernBERT-large achieved the greatest performance gains, with the improvements in validation F1 scores of 1.0% and 2.5% at the division level, 4.4% and 2.2% at the group level, and 9.9% and 11.5% at the field level. An emergent ability was observed, as performance in non-augmented classes improved with ModernBERT-large but not with ModernBERT-base. Overall, this study demonstrates that class mapping effectively enriches training datasets, enhances classification performance, and underscores the importance of model size and architecture. These findings offer a practical and scalable strategy for improving machine learning performance in research classification tasks.
The metaverse is a connected virtual world blending digital and physical spaces, allowing people to interact as they do in real life. This study employed a 2D metaverse-based digital curation, designed to allow learners to experience the historical event of “the internment of Taiwanese civilians in Hong Kong during World War II,” to provide interactive and immersive history learning activities. This approach includes digital narratives and environments, promoting collaborative learning and a deeper historical understanding from a first-person perspective. To test the effectiveness of this learning mode, 17 Grade 10 students from Taoyuan City, Taiwan, participated in a single-group pre-experimental design. The study aimed to examine whether this learning mode improved historical learning and how it affected students with different prior knowledge and cognitive styles (verbal or visual preferences). Behavioral analysis identified effective learning patterns in the metaverse. The results showed significant improvement in learning performance, especially for students with lower prior knowledge and a visual learning style, across memory, comprehension, and critical thinking. High-performing students explored the exhibits thoroughly and revisited them frequently, enhancing their learning. Interviews revealed that students found digital curation more engaging and memorable than traditional methods, thanks to its interactive elements and rich narratives. The study suggested specific strategies for integrating metaverse digital curation into history education. While the metaverse is still developing, its potential in education is significant and warrants further research.
During the past, COVID-19 pandemic has significantly impacted global development in various respects. The extent of its effects on individuals depends on the preventive and healthcare actions they take. Therefore, the quality of information they encounter plays a crucial role in navigating the pandemic. Misinformation about COVID-19 has permeated society, spreading through diverse social media. College students, who are common users of social media, are particularly vulnerable to misinformation. Consequently, students have a higher risk of exposure to misinformation. The purposes of this study were to examine how college students experienced and behaved in response to COVID-19 misinformation disseminated through social media and the effort needed for prevention. A survey approach was employed to collect data from 300 respondents. In addition, in-depth interviews of 10 college students and 3 infectious disease physicians in hospital were also conducted. From the results of the study, policies, training, and technology-based interventions were suggested to prevent misinformation. Joint efforts from society are needed to prevent the spread of health-related misinformation. College students also need to be aware of the sources of reliable resources to access health-related information, and need to develop appropriate health-information literacy, understand the nature of infectious diseases, and avoid the health risks caused by misinformation.
Lotus petals were once inserted into book pages by our predecessors as a method of preventing bookworm damage in ancient times, a topic on which scholars have conducted little research and rarely provided practical examples. By combining documentary records with historical practices and utilizing sources such as poetry collections, archives, diaries, memoirs, and newspapers, the history of using lotus petals for bookworm prevention can be traced. This research aims to authenticate the origins of this method and explore its regional distribution characteristics. Temporal distribution can be illustrated through poems and inscriptions on lotus petal papers by Qing dynasty literati, suggesting that the practice of inserting lotus petals into books was already in use during the Qing dynasty, with explicit textual records appearing from the late Qing period onward. The use of lotus petals for bookworm prevention became common knowledge from the late Qing dynasty to the early Republic of China period. Regarding geographical distribution, this practice was limited to regions where lotus flowers grew abundantly and was only employed when the petals were fresh, typically during the summer when lotus flowers were in full bloom.
Shan Dong Jun Xing Ji Lue ((sic)(sic)(sic)(sic)(sic)(sic)) is a precious historical material for studying the military operations of the Taiping Army ((sic)(sic)(sic)), the Nein Army ((sic)(sic)), the White Lotus Society and another uprising army in Shandong. However, there is no clear understanding of the authorship, edition, writing, publication, and distribution of this book. By comparing the existing manuscripts of the Shan Dong Jun Xing Ji Lue with some other editions, a clear conclusion can be drawn from the above questions. Firstly, this book was solely written by Guan Yan ((sic)(sic)), an aide of Yan Jing-Ming ((sic)(sic)(sic)). Secondly, Zhang Yao ((sic)(sic)), the author of the postscript of Shan Dong Jun Xing Ji Lue, shares the same name with the Shandong governor Zhang Yao, but they are not the same person. Finally, the manuscript of the Shan Dong Jun Xing Ji Lue was originally completed in 1874, later published in 1876, printed in lead type by the Shen Newspaper Office ((sic)(sic)(sic)) in 1879, and reprinted in 1885 with the addition of Zhang Yao's postscript. According to the characteristics of the book editions, the order of the editions existing today can be confirmed.
This study conducted a questionnaire survey targeting librarians in municipal public libraries in Taiwan to examine their perception of job rotation. The aim was to identify the relationships among job rotation perception, job characteristics perception, job satisfaction, and intention to stay. It also sought to explore whether implementing job rotation in public libraries impacts librarians' job characteristics perception, job satisfaction, and intention to stay. The results of the analysis revealed that job rotation perception, job characteristics perception, and job satisfaction all have a significant positive effect on intention to stay. Job rotation perception positively influences both job characteristics perception and job satisfaction. Job satisfaction mediates the relationship between job rotation perception and intention to stay, but perceived job characteristics do not. Based on the research findings, the authors propose four suggestions: (1) improving the job rotation system; (2) reinforcing the importance of library work; (3) enhancing external incentive measures; and (4) increasing the significance of job rotation.
Introduction. Notwithstanding the growing interest in employing electroencephalography (EEG) for web search research, studies presently being conducted continue to grapple with significant issues, namely, small sample sizes, variability in methodology, and limited generalisability. This systematic review seeks to address these particular issues by describing how EEG has been used in this field of research, addressing sample sizes, active authors, methodologies, and limitations. Method. This systematic review employs the PRISMA framework to analyse the application of EEG in web search studies, focusing on sample sizes, methodologies, and challenges faced by researchers. A comprehensive search was conducted across multiple academic databases to identify relevant studies. Results. Findings indicate that typical sample sizes in EEG studies range from 10 to 24 participants, largely due to resource constraints. Researchers encounter challenges such as biological artefacts affecting data quality, the complexity of emotional and cognitive states, and limitations in generalisability due to small sample sizes. Additionally, issues related to equipment quality and methodological consistency further complicate EEG research in this domain. Conclusions. The application of EEG in web search research holds significant potential for enhancing our understanding of user interactions with search engines. However, addressing the identified challenges is crucial for improving the robustness and applicability of findings. Future studies should focus on refining methodologies and exploring innovative approaches to overcome existing limitations in EEG research related to web searches.
Background: The COVID-19 pandemic increased anxiety and depression rates, prompting heightened research activity in mental health, particularly among vulnerable populations like university students. Research activity adapted in order to address issues arising in this new landscape. Objectives: This study aims to explore the changes in scientific activity and response trends during a public health crisis, with a particular focus on analyzing research themes, categories, and collaboration patterns related to the mental health of Chinese university students before and after the COVID-19 pandemic. Method: Using bibliometric methods, data were extracted from Web of Science to analyze scientific output related to the mental health of Chinese university students from 2017 to 2022. Descriptive statistics and visual analysis tools were employed to explore publication trends and collaboration patterns. Discussion: After the pandemic, there is an increase in the total amount of literature and open-access publications. Research topics have shifted towards prioritizing student well-being over clinical diagnoses. Strengthened international and institutional collaboration is evident through increased cooperation with developing countries and those heavily impacted by the pandemic, as well as intensified partnerships between medical universities and prestigious academic institutions. Conclusions: The findings of this research can guide supportive efforts by libraries, academic journals, international partnerships, and funding bodies to address mental health challenges during future epidemics. Furthermore, they highlight the value of collaboration between bibliometrics and psychology in tackling the complex issues surrounding mental health in such contexts. Professionals in bibliometrics and psychology should collaborate to address the complex challenges posed by mental health and future epidemics.
This scoping review systematically examines the evolution of research impact assessment frameworks, categorising their development into 4 generations. The study follows Arksey and O’Malley’s (2005) scoping review methodology, enhanced by Levac et al. (2010) and aligned with PRISMA-ScR guidelines. A comprehensive search was conducted across Scopus, Web of Science, PubMed, IEEE Xplore, and Google Scholar, as well as grey literature, to identify relevant studies on research impact frameworks, metrics, and methodologies. A total of 139 studies were selected based on predefined inclusion criteria, encompassing bibliometric indices, multidimensional frameworks, predictive analytics, and alternative metrics. The findings highlight the transition from traditional citation-based measures to sophisticated, data-driven methodologies. The first generation (bibliometric indices) included 26 key metrics focusing on publication productivity. The second generation introduced 41 multidimensional frameworks incorporating societal, economic, and policy indicators. The third-generation integrated machine learning and predictive analytics to assess impact across 17 aspects and 9 data-driven factors. The fourth generation utilised 16 alternative metrics, including Altmetric Attention Scores and PlumX Metrics, to capture real-time digital engagement. Key challenges identified include limited standardisation, regional biases, and the under-exploitation of emerging technologies, such as large language models. The study underscores the need for predictive multidimensional frameworks and standardised taxonomies to enhance scalability and foresight in impact assessment. By structuring the evolution of research impact science, this review provides actionable insights to refine assessment methodologies, ensuring their relevance for addressing societal challenges and guiding strategic research investments.
The Lu Zhuan ((sic) Ba ((sic) Du Gong Bu Ji ((sic) chao ben ((sic) in Fu Ssu-nien Library ((sic) is a rare book. Comparison with Qian Zhu Du Shi ((sic), engraved by Ji Zhen-Yi ((sic) in the sixth year of the Kangxi((sic) reign (1667), reveals obvious similarities and differences between the two books, and leads to the speculation that this book was an intermediate draft before the finalization of Qian Jian ((sic). It preserves important information about the bian ci ((sic) and jiao yu ((sic) of "Wu Ruo Ben" Du Gong Bu Ji ("(sic), and is the closest edition of Wu Ruo Ben. Wu Ruo Ben is the second most important Du ji ((sic) after the zu ben ((sic) Er Wang Ben ((sic)), but it has been lost. The question of the appearance of the Wu Ruo Ben is the key to the study for early Du ji. The introduction of Fu Tu Ben ((sic) can provide new material for solving this problem. Combining Fu Tu Ben and the related Du ji can make general conjectures about Wu Ruo Ben's bian ci, inclusion of poems and articles, annotations, texts and jiao yu. The examination of Fu Tu Ben can advance the study of the Wu Ruo Ben, and is also of great significance to the study of early Du ji such as Er Wang Ben.
The Yu Tai Xin Yong ((sic)(sic)(sic)(sic)) edition printed by Zheng Xuan-Fu ((sic)(sic)(sic)) in the Jiajing period ((sic)(sic)) in the 19th year of the Ming Dynasty ((sic)(sic)) is the foundational edition in Zheng Xuan-Fu editions ((sic)(sic)(sic)(sic)(sic)), one of the two major editions of Yu Tai Xin Yong. Zheng's edition of the Yu Tai Xin Yong was compiled and edited by Zheng Xuan-Fu based on an incomplete edition and a handwritten copy, differing from the other major editions of Yu Tai Xin Yong, Chen Yu-Fu editions ((sic)(sic)(sic)(sic)(sic)), therefore contains information not present in the Chen editions and holds unique textual value. Another significant aspect of Zheng's edition is that it reintroduced Yu Tai Xin Yong to the public eye, greatly influencing its transmission. The editions circulated during the Ming Dynasty primarily belong to Zheng's editions. However, there is considerable confusion and error in the current records of ancient books that the supposedly Jiajing period Xu Xue-Mo ((sic)(sic)(sic))'s Hai Shu Lou ((sic)(sic)(sic)) edition is actually a Wanli ((sic)(sic)) period reprint that omits information about Zheng's original carving, which does not affect Zheng's edition's status as the original and authoritative Ming edition. Additionally, the so-called Jiajing 22nd year Yang Shi-Kai ((sic)(sic)(sic))'s edition, claiming to be based on a Song Dynasty edition ((sic)(sic)), is actually a recarved ((sic)(sic)) print derived from Xu's edition, with no genuine connection to the Song source.
