[Purpose/Significance] Clarifying the value co-creation mechanism of open public data is of great significance for unleashing the value of open public data and creating a value co-creation ecosystem. [Method/Process] Through literature research, Mitchell scoring method, relevant stakeholders were identified and classified. Then, from the perspective of supply and demand matching, the value behavior and relationships of each stakeholder were analyzed. Based on this, a two-dimensional analysis framework of “horizontal implementation process — vertical collaboration process” was constructed to reveal the value co-creation mechanism. [Result/Conclusion] The study finds that there are 10 stakeholders involved in the value co-creation of open public data, which can be divided into three categories based on their level of participation: core, sub core, and peripheral stakeholders. From the perspective of horizontal implementation process, each stakeholder undergoes activation, aggregation, and diffusion of data value to achieve value co-creation. From the perspective of vertical collaboration level, each stakeholder goes through the service exchange layer, resource interaction layer, and reaches the value co-creation layer.

[Purpose/Significance] This paper explores the key influencing factors and mechanism of algorithm literacy on users’ adoption behavior of recommended information, aiming to provide a theoretical basis and practical guidance for improving the user experience and user adoption of algorithm recommendation. [Method/Process] Based on the FATE model and self-efficacy theory, a theoretical model of the influence of algorithm literacy on users’ adoption behavior of recommended information was constructed, and empirical analyses were conducted with SEM and fsQCA. [Result/Conclusion] The SEM results indicate that algorithm perception and its sub-dimensions, algorithm skills, and algorithm literacy self-efficacy have a significant influence on users’ adoption behavior of recommended information, while the algorithm knowledge does not. The fsQCA analysis reveals 8 configurations that positively affect users’ adoption behavior, indicating that the constituent elements of algorithm literacy need to work synergistically to be effective. Among them, perceived algorithmic transparency, perceived algorithmic fairness, algorithm skills, algorithm knowledge, algorithm literacy self-efficacy, and adoption attitude are the key factors contributing to users’ adoption behavior.

[Purpose/Significance] Evaluating the maturity level of smart library development in universities is crucial for the self-reflection on the effectiveness of smart library and the formulation of future development plans. [Method/Process] This paper designed a sampling questionnaire based on the maturity evaluation model of smart library development of universities. It selected 21 public universities in Beijing to obtain index data through questionnaire survey, and carried out an empirical evaluation of smart library development of universities. [Result/Conclusion] The development of smart libraries in universities in the capital is in the developing stage. Each library presents different characteristics in the dimensions of librarian team, infrastructure, business management and user service. According to the results, the paper puts forward the optimization suggestions for specific construction contents in each dimension from the micro perspective, and proposes maturity level improvement strategies from the initial level to the optimization level from the macro perspective.

[Purpose/Significance] The knowledge element database is constructed by integrating the potential multidimensional relationship between South China Sea rights protection evidence, which can represent the massive evidence and its association in a deep and all-round way. In turn, it provides large-scale, high-quality, and fine-grained annotated corpus to support the construction of an evidence chain for South China Sea rights protection. [Method/Process] First, considering the existing knowledge element database and based on the knowledge element representation and linking theories, facing the needs of evidence chain construction, this paper proposed a theoretical framework for an evidence knowledge element database for South China Sea rights protection and deeply explained the core content of evidence knowledge element representation and evidence knowledge element association. Second, based on this theoretical model, the construction path of the knowledge element database was proposed. The evidence knowledge elements were modeled in stages from shallow to deep in the carrier layer, content layer, value layer, and association layer, and integrated into a knowledge element database through the four sub-databases. Third, some text and map evidence knowledge elements were selected to verify the feasibility of the proposed theoretical framework and realization path. [Result/Conclusion] It can provide a reference for the relevant research on the South China Sea document collection, knowledge element database, and the evidence chain for South China Sea rights protection. The knowledge element database for scaled multidimensional South China Sea rights protection evidence needs further research and implementation.

[Purpose/Significance] In the context of interdisciplinary integration, analyzing and grasping the current situation and future direction of interdisciplinary output influence of disciplinary knowledge is conducive to promoting the development of discipline and improving the academic discourse power. [Method/Process] The interdisciplinary citation sentence is the intermediary “bridge” connecting the output knowledge with the received knowledge. On one hand, their semantic expression can reflect the closeness of the relationship between the two disciplines. On the other hand, their location and emotional tendency can reflect the different stimulating effect and the discipline knowledge value on interdisciplinary knowledge innovation. This paper constructs an influence model of a target discipline knowledge on interdisciplinary knowledge from two aspects: knowledge relevance and knowledge innovation efficiency. The output of one knowledge has an impact on multiple interdisciplinary knowledge, so the overall influence is calculated by summing, simultaneously identifying the output knowledge of the target discipline and the influenced interdisciplinary knowledge. [Result/Conclusion] Taking the data of Information Resource Management discipline from 2017 to 2022 as an example, the validity of the model is verified. It finds that knowledge of this discipline mainly influences the knowledge innovation in social sciences, with less impact on natural sciences. The influence of mature cross-knowledge is stronger than that of professional core knowledge. It mainly affects interdisciplinary knowledge, being less effective in influencing the core knowledge innovation of other disciplines.

[Purpose/Significance] This study aims to improve automatic sentence segmentation and punctuation in ancient Chinese texts using generative large language models, overcoming the limitations of traditional sequence labeling models that require specialized label design and annotated data construction. [Method/Process] In order to increase the number of learnable samples, a sliding window strategy was used to segment training data. Additionally, MinHash and Locality-Sensitive Hashing (LSH) were employed to provide reference samples for unpunctuated texts, and constraints were applied to the decoding process of the large language model. The Xunzi ancient literature large language model served as the base model. Fine-tuning was performed using Low-Rank Adaptation (LoRA) to align the model with the task of ancient Chinese automatic punctuation, enabling the generation of target texts with punctuation from unpunctuated texts. [Result/Conclusion] On two benchmark datasets released by EvaHan 2024, the F1 scores of automatic sentence segmentation are 88.47% and 92.48%, and the F1 scores of automatic punctuation are 75.29% and 80.01%, respectively. These results significantly outperform the Xunzi ancient large language model and ChatGPT 3.5, indicating that generative large language models are a feasible approach for addressing the tasks of sentence segmentation and punctuation in ancient Chinese texts.

[Purpose/Significance] The study of the health information behavior of sport groups on the content social platform not only improves the cognition and literacy of personal health information, but also expands the group boundary of research. It can also provide theoretical reference for personalized recommendation services on the content social platform. [Method/Process] Based on the S-O-R theory, this study constructed a logical framework and adopted web crawler technology to collect the text data from sports records on Xiaohongshu. Then, it analyzed the thematic characteristics of user health information behavior by the sentiment classification from sentiment lexicons and K-means clustering methods from machine learning. [Result/Conclusion] Based on the S-O-R theory, the thematic characteristics model of user behavior is constructed based on the subject words. The model provides a research basis for users to improve their awareness of physical activity and institutions to launch high-quality health information services.

[Purpose/Significance] The accessibility of public digital culture services links supply and demand, and it essentially reflects the user’s evaluation of the degree of match between supply and demand. Measuring and exploring users’ evaluation and formation mechanism of the accessibility of public digital culture services is helpful for understanding the situation and problems and formulating continuous optimization countermeasures. [Method/Process] This article used a mixed research method that contains the fuzzy comprehensive evaluation method to quantitatively analyze 442 valid questionnaires from a field survey in Chengdu, and grounded theory to construct a theoretical framework for the formation mechanism of accessibility users’ evaluation. [Result/Conclusion] The users’ evaluation of the accessibility of public digital culture services in Chengdu is generally positive, but availability and adaptability are still insufficient. The five dimensions of the accessibility of public digital culture services are presented as a “cone” structure, with spatial accessibility at the highest “cone point” position, acceptability and awareness at the bottom, and accessibility and adaptability in the middle. The users’ evaluation of the accessibility of public digital culture services is formed in the process of inter-embedding supply and demand, in which resource allocation and service quality affect the outcomes of supply, and users have to cross the barriers of individual and social, organizational and economic and technological factors, which confirms the classical framework of “the formation mechanism of medical and health service accessibility across barriers” advocated by M. Gulliford. However, due to the differences in the basic public service supply system between China and the West and specific service types, the formation mechanism of users’ evaluation of public digital culture service accessibility has local heterogeneous.

[Purpose/Significance] Disinformation refers to the atmosphere of uncertain information that influences or misleads others in making decisions by disseminating information that is intertwined with truth and falsehood. By quantitatively analyzing EU disinformation reports, this article aims to explore the research themes and focuses, summarize governance experiences and obstacles. Finally, based on China’s national conditions and EU’s governance practice, it proposes a Chinese style approach to deal with disinformation. [Method/Process] Using policy reports from various EU institutions on disinformation from 2018 to present, released on the website of EU Publications Office, as the core corpus, and the solution developed by the Carnegie Endowment for International Peace for EU disinformation as a supplement, the LDA topic model was used to calculate topic popularity and threshold, analyze hot topics, and explain the evolution of topics. [Result/Conclusion] Through the LDA model, a total of 8 EU disinformation related topics are excavated, and 3 hot topics are identified. The study finds that the EU disinformation policy and governance begins with the definition of disinformation. The research focus gradually shifts to the audience of disinformation and democracy and freedom issues in the technology governance. At the same time, measures such as changing platform advertising revenue channels, supporting factual verification work, and popularizing media literacy education are explored, achieving good governance results. However, there are also limitations in passive governance, lack of data access plans, and specialized resources.

[Purpose/Significance] To provide reference for named entity recognition in biomedical field, this article compares the effect of the discriminative language model represented by BERT and the generative language model represented by ChatGPT in Chinese medical texts named entity recognition tasks. [Method/Process] This article designed six models of BERT splicing different types of deep learning networks. Meanwhile, prompts with different dimensional elements were designed for ChatGPT, and the extracting results were evaluated by exact matching and loose matching respectively. [Result/Conclusion] Among the discriminative language models, BERT-BiLSTM-CRF model has the best performance, with F1 values of exact match and loose match reaching 85.71% and 89.49% respectively. In generative language models, GPT_baseline+A+three-shot model that combined with prompt language framework achieves the best results, and its exact match and loose match F1 values are 78.14% and 84.73% respectively. In this study, BERT series model achieves relatively best results. The well-designed prompt framework for ChatGPT could significantly enhance the model’s understanding of named entity recognition tasks, and is expected to be a new breakthrough in natural language processing tasks in the future.

[Purpose/Significance] Zhang Taiyan’s QilueBieluYiwenzheng is a family collection of manuscripts. The Complete Works of Zhang Taiyan, published by Shanghai People’s Publishing House in 1982 and 2014, include a “collated and punctuated edition” of this text. However, this edition contains numerous punctuation errors that deserve focused correction. [Method/Process] This study carefully examines the original text, refers to records in related ancient books, and corrects 29 punctuation errors in the “collated and annotated edition” of QilueBieluYiwenzheng. [Result/Conclusion] The aim of QilueBieluYiwenzheng is to collect the lost writings from Liu Xiang and Liu Xin’s Bielu and Qilue. This work engages with the general theories and methods of the philology and the principles and paradigms of textual criticism and bibliography. The 29 represent common mistakes in punctuating classical Chinese texts and also reflect a relative lack of knowledge in the study of philology and textual criticism and bibliography.

[Purpose/Significance] This study achieves the protection and adaptive use of the cultural heritage of Manchu ancient books by constructing a knowledge base of these books. It serves as a reference for protecting other minority heritages and promotes the creative transformation and innovative development of China’s outstanding traditional culture. [Method/Process] This study structured the resources of Manchu ancient books. It constructed six types of ontology models of offices, people, places, organizations, time and events, and defined entity attributes and relationships between entities. Using the ALBERT-BiLSTM-CRF model, it extracted entities and relationships to form a knowledge map and further constructed a knowledge base platform. [Result/Conclusion] Based on the knowledge base platform of Manchu ancient books, it achieves knowledge retrieval, data analysis, intelligent recommendation, AI dialogue and other knowledge services for Manchu ancient books, and improves the efficiency and effectiveness of knowledge services for ancient books.

[Purpose/Significance] This paper systematically reviews and analyzes the related research and practice of research reports in humanities and social sciences, aiming to provide reference and inspiration for strengthening the construction of a diversified evaluation system and the classification evaluation. [Method/Process] Firstly, it sorted out the concepts, attributes and categories of the research reports achievements in humanities and social sciences. Secondly, it started from the necessity of evaluating the reports, and summarized the observation points and methodology of the evaluation. Lastly, it conducted some discussion and reflection on existing research. [Result/Conclusion] At present, the concept of research report has not been unified yet. The evaluation observation points need to be further optimized, and the evaluation method becomes more comprehensive. In terms of evaluation objects, it recommends to clarify the concept and attribute characteristics of research report. In terms of evaluation subjects, it is necessary to identify the needs of different subjects and set evaluation observation points from multiple perspectives. In terms of evaluation methods, it recommends to improve the comprehensive evaluation method and promote the evaluation database construction for research report results.