Page 11 - JOURNAL OF LIBRARY SCIENCE IN CHINA 2018 Vol. 42

010 Journal of Library Science in China, Vol. 8, 2016

of database resources, digital libraries should also include government information, society-
concerned information, Internet hot topics, as well as user behaviors like use of digital libraries and
network search, as part of their data collections. Moreover, digital libraries should also increase the
production, collection and storage of renewable resources, including the various types of integrated
data after correlation, mining and analysis, and finally diversify their data resource categories.
(2) Spread the resource integration of digital libraries
The unique feature of big data is the complexity and diversity of data. Pure literature information
could not meet the needs of digital library users, which inevitably leads to the conversion of
resource integration from pure literature service to comprehensive information service. In order to
meet various needs of users, we need to integrate literature information, government information,
social information, network information, and user information. In other words, we need to
associate digital library resources with social resources, especially the social resources of public
concern.
In addition, we also need to strengthen the resource integration oriented to complex problem
solving. In a big data environment, users expect digital libraries to help or guide them solving
practical and complex problems instead of simply meeting their demands related to literatures.
Therefore, the traditional resource integration based on simple accumulation and classification
is confronting great challenge. To meet the actual needs of users, we must carry out the resource
integration oriented to complex problem solving.
(3) Deepen the resource organization and processing of digital libraries
Resource organization and processing arrange the sequence of resources, coming up with a
presentation of resources in the form of knowledge. In digital libraries, structured information (e.g.
bibliographic information) itself has certain forms of knowledge presentation, which is basically
restricted within the repository. In a big data environment, we must widen the views, establish a
semantic link between internal and external resources of digital libraries, and construct a complete
social resource oriented digital library resource organization architecture.
In addition, for the purpose of further deepening the processing of library resources, we must
upgrade from deconstructive processing of information to constructive processing of information.
For example, we can break down the literature into knowledge units (knowledge deconstruction),
establish semantic association among these knowledge units or between these units and the
literature (knowledge construction), and then produce new knowledge. In the deconstruction and
construction process of knowledge, we should make use of data mining technology and analysis
software, and improve the capability to apply and share the produced subject headings and
classification charts.
In the resource organization and processing of digital libraries, another non-literature resource
should draw our attention, which is the user behavior information. User behavior information is a
precious resource in the knowledge service and knowledge recommendation of digital libraries and
requires deep exploration and proper organization. If the user behavior data is closely associated