[Purpose/Significance] Knowledge discovery of oral history archives resources based on knowledge map is a new attempt of knowledge discovery in the field of digital humanities, which provides a new path for fine-grained association, semantic query and personalized exploration of resources.[Method/Process] Taking the data of the Anti-Japanese War Veterans Oral Data Center of Nanjing Normal University as the data source, the knowledge map of oral history archives resources of anti-Japanese war veterans was constructed. Based on the map examples, multidimensional knowledge discovery research was carried out from the aspects of overall project overview, event theme relationship, social network relationship, space-time network relationship and so on.[Result/Conclusion] Digital human technology represented by knowledge map provides powerful tool support for knowledge discovery research and injects new kinetic energy into the deep development of human resources.

[Purpose/Significance] Combing the historical facts with storyline not only has a certain theoretical and practical guidance and significance for describing and grasping the direction of historical development, but also provides a new technology realization mode and innovative research perspective for knowledge discovery in the humanities field.[Method/Process] This study provided a research paradigm of storyline construction and visualization, emotion and scene based on text data, and made empirical exploration by combing the text of Zhang Xueliang's Oral History as the data source. This paper eused jieba word segmentation to clean the initial data source of Zhang Xueliang's Oral History to obtain experimental data source. LDA topic model was used to obtain topic distribution and t-SNE data dimension reduction was performed to present topic module. With the help of SnowNLP emotion dictionary, emotional feature words were mined, Zhang Xueliang's emotional evolution stage was sorted out, and the storyline was constructed.[Result/Conclusion] Through the construction of Zhang Xueliang's storyline, the dynamic mutual promotion of multi-dimensional elements such as characters, places, events, time and emotions is realized.

[Purpose/Significance] Based on the perspective of digital humanities, through the visual display and analysis of the time-space information of relegated poets in Complete Collection of Tang Poetry, it can become the research evidence in the field of history and literature, help related scholars find new research clues and provide references for them.[Method/Process] Using the method of measurement and statistics and the QGIS geographic information system, through the collection, cleaning, visualization and analysis of the data, this paper visualized and analyzed the time-space information of the relegated poets in Complete Collection of Tang Poetry.[Result/Conclusion] The visual research on the time-space trajectory of relegated poets in Complete Collection of Tang Poetry is helpful to explore the mining methods of related elements of relegated poets, and provides new clues for the study of Complete Collection of Tang Poetry in the field of historical literature.

[Purpose/Significance] This paper designs a scientific and standardized fine-grained semantic description model of modern newspaper resources, and reveals the characteristics and relationships of modern newspaper resources in depth, in order to provide references for the effective management, organization, knowledge discovery and knowledge service of modern newspaper resources.[Method/Process] By analyzing the logical structure, physical layout, contents of modern newspaper resources, starting from two aspects of domain ontology description and metadata description, this paper reused the CIDOC-CRM ontology conceptual model, EAD, DC and Ancient Book Metadata Specification, and took Shengjing Times as an example to design a fine-grained semantic description model for modern newspaper resources. Then, it used Oxygen XML tool to describe the semantic model in RDF/XML language to realize element interoperability and model application.[Result/Conclusion] This paper aims to provide an operable fine-grained semantic description model for the modern newspaper resource organization, provide a basic guarantee for the construction of the modern newspaper resource database, the standardized newspaper management and the application system development, and promote the development, utilization and sharing of the modern newspaper resources.