Page 114 - JOURNAL OF LIBRARY SCIENCE IN CHINA 2015 Vol. 41

Ying WANG, Zhixiong ZHANG, Hui SUN & Feng LEI / A method of semantic representation and 113
organization of the historic knowledge on contemporary China

,OPXMFEHF 0SHBOJ[BUJPO
/BWJHBUJPO #SPXTF $ISPOJDMF PG &WFOUT $ISPOPMPHZ

/FX 0SHBOJ[BUJPO 4DIFNB
5FYU 0SHBOJ[BUJPO

&WFOU $POGFSFODF
… UJNF EJNFOTJPO
UZQF EJNFOTJPO
*UFN -FWFM
0SHBOJ[JOH VQ
5FYU *UFN *OGPSNBUJPO 5FYU *UFN 5FYU *UFN
3FTPVSDFT
'BDU -FWFM
'BDU
'BDU PCK SFMBUJPO PCK 'BDU
PCK SFMBUJPO PCK PCK QSPQFSUZ WBMVF

0CKFDU -FWFM
.JOJOH EPXO
PCKFDU PCKFDU PCKFDU
,OPXMFEHF
&YQMPSBUJPO
3FUSJFWBM "TTPDJBUJPO 3BOL $MVTUFSJOH3FPSHBOJ[BUJPO

/FX 0SHBOJ[BUJPO 4DIFNB
Figure 1. Framework of the method of semantic representation and organization.

a historical knowledge network. By using text mining technologies, this process identifies
knowledge objects from unstructured texts, annotates the important sentences that contain these
objects, and then extracts the related facts which form important structural facts represented in
“object1-relation-object2” or “object1-property-value” based on the judgment of historical
experts. Many facts about one knowledge object may be extracted from the same or different text
items and different objects may be associated directly or indirectly. Moreover, a text item will be
related to others according to the same object or fact. In this way, the historical knowledge network
can be constructed according to the association of knowledge objects, facts and text items. Based
on the network, historical knowledge can be further explored, such as retrieval, association, rank,
clustering, and reorganization and so on.
“Organizing up” is a construction process, which utilizes the historical knowledge network
constructed by the “Mining down” process to implement deeper level organization and applications
on historical knowledge and text items by associating or reorganizing them. Currently, the
historical text items on contemporary China are simply organized by time dimension (such as
“1949.09-1956.09” period, “1956.10-1966.04” period, etc.) and type dimension (such as event,
conference, organization, etc.). The knowledge hidden in text still cannot be effectively discovered.