Page 107 - JOURNAL OF LIBRARY SCIENCE IN CHINA 2018 Vol. 42

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

information. Moreover, thesaurus metadata service reveals a holistic overview of Chinese thesaurus
construction, as well as the features of thesauri in disciplines, construction and structure. Thus,
users can choose appropriate thesauri according to individual requirements. Thesaurus metadata
service can be considered as a set of most basic data services, which facilitates resource discovery
and the utilization of thesauri.

2.3.2 Concept and concept group service
Concept and concept group are knowledge units of different granularities in thesaurus knowledge
systems. Concepts are the smallest description units in thesaurus knowledge systems. Concept
groups are knowledge units with multiple concepts interconnected according to specific logic. For
example, a group consisted of a concept and its connected (broader, narrower, related) concepts,
or a group consisted of concepts connected through hierarchical relations in a concept hierarchical
tree, or a group consisted of concepts from multiple thesauri connected through mapping relations,
or a group consisted of concepts in same category or ontology class, and so on. Concept service
has several levels, according to the completeness of provided concept description information.
For example, only Chi-Eng labels of concepts are provided, or only synonyms are provided, or
hierarchical relations and association relations are provided simultaneously, or cross-thesauri
concept mapping relations are provided, and so on. Service mode of concepts and concept groups
should be designed and implemented step by step, according to demands on knowledge units in
thesaurus applications.

2.3.3 Thesaurus customization service
More than 100 thesauri have been constructed in China. Although these thesauri cover almost all
disciplines, they cannot fulfill all demands from the real practice. There are several main reasons:
1) many thesauri are relatively old. Most thesauri were constructed in the 1990s and have not been
maintained properly. Therefore, a number of new concepts have not been included; 2) traditional
thesauri were designed to meet the needs of artificial indexing and have clear characteristics of
artificial languages. Terms in these thesauri cannot properly match with the terms in articles. Many
special concepts need to be obtained through artificial concept coordination. Hence, it’s difficult
for these thesauri to meet the demands on computer information processing; 3) traditional thesauri
usually focus on one or several disciplines. The width of disciplines and the depth of knowledge
cannot be covered at the same time. Therefore, it is difficult to meet demands on information
organization and reveal in specific subject information systems. In addition, as formal knowledge
systems, thesauri can be used not only in information organization, but also in knowledge mining,
knowledge discovery, etc. However, it’s difficult for existing thesauri to support these utilizations.
Therefore, thesaurus customization service that provides data and tools to support new thesaurus
construction is important for NTW.
Thesaurus construction based on NTW requires the following steps: 1) choosing terms about a