Page 169 - Journal of Library Science in China, Vol.47, 2021

168 Journal of Library Science in China, Vol.13, 2021

archivists to supplement and save the context content of archives into the corresponding database
under the premise of possessing archival materials and knowing the results of object extraction,
combined with the analysis results brought by digital methods, focusing on creating events and
spatial environments related to archives or archives bond itself, so as to maintain the inherent value
of archives before deconstruction.

3.2 Archival data organization technology from the perspective of value mining

The perspective of value mining closely corresponds to the “reorganizing” session in the archival
data research methodology, aiming to describe, connect and aggregate archival data through
multi-dimensional knowledge organization models and technologies, forming a dynamic archival
semantic knowledge graph from DH perspective, and further exploring its historical value and
cultural value on the basis of previous voucher value and reference value, to form a clear “value-
added” path of archival data. The memory entities, entity semantic relationships and context
content such as people, time, and place obtained through the “discovering” session are stored in
different types of databases in a weak correlation, and uniformly linked to the original text of the
archival data, and the position and arrangement relationship of these elements anchored in the
original text are still recorded.
As mentioned before, archival data organization is divided into two steps: “Static association”
and “Dynamic aggregation”. “Static association” standardizes and organizes the extracted
instances and their semantic relationships through the ontology model, forms the basic framework
of the knowledge graph of archival data, imports it into the graph database for storage, and
establishes a semantic retrieval and question answering mechanism for archival knowledge
that integrates knowledge graph technology. “Dynamic aggregation” focuses on the thematic
clustering and division of entities in the underlying resources, and obtains the semantic and
contextual similarity between entities through knowledge calculation, which will gather around
the descriptive concepts of the same entity. At the same time, with the help of dynamic knowledge
organization models with an internal “entity type” (sem: EntityType) conceptual system such as
“Event Ontology” or “Simple Event Model” (SEM) (such as “event” (sem:Event) concept has
“event type” (sem:EventType)), the cognitive dimension of the entity is extended and cut in from
multiple aspects. And use the canonical classification system to organize conceptual examples in
[27]
multiple dimensions .
Based on the SEM model, this paper takes the relationship between Селезнев (the former Soviet
archivist) and WU Baokang (the founder of the Department of Archives of Renmin University of
China), as an example to construct a knowledge organization model based on multidimensional
cognition. In Figure 6, the semantic relationship “teacher-student relationship” between the two
instance nodes “Селезнев” and “WU Baokang” appears in the knowledge graph as an empty
node, acting as a specific role in the event, and its role type is regulated by the Chinese Library