Page 147 - JOURNAL OF LIBRARY SCIENCE IN CHINA 2018 Vol. 42

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

we can extract method terms from academic papers. This information can be used as a key data
source for constructing the method knowledge base. The method knowledge base can be embedded
into decision support decision system or expert intelligence system and act as a support for method
selection. We can also use the method knowledge base to draw method pedigree chart or field
method development map. These results will promote the normalization of the usage of methods.
In a word, the extraction and mining of method knowledge elements have both research and
application value.
Knowledge mining can be categorized into two approaches, i.e. the statistical-based and the
rule-based. Statistical-based approaches are more suitable in text-mining fields that the unit of
processing is the word. However, in the fields that the unit of processing is the sentence, due to the
complexity of sentences, it is difficult for statistical-based approaches to adapt to such a situation
and thus researchers prefer rule-based and pattern recognition approaches. The description of
method knowledge elements in academic papers is sentence-based. It is more appropriate that we
choose rule-based and pattern recognition approaches. The types of the method knowledge element
and the way to construct method knowledge element description rules are the key in the extraction
of method knowledge element from papers.

1 Literature review

Current studies on knowledge elements and rule extraction focus on several aspects, e.g. theories
and reviews, knowledge element representation and modeling, knowledge element and rule
extraction and implementation, etc.
(1) Knowledge element theories and reviews
Wen Youkui et al. (2011) believed that the knowledge element semantic linking theory would
be important in the next-generation knowledge discovery pattern. He explained extensively on the
basic concepts, methods and technology of discovery of latent linking between knowledge elements
in literature from a semantic linking theory perspective. Gao Jiping et al. (2015) presented a review
of the definition, measures, research projects and applications of knowledge elements on various
fields. Jiang Yongchang (2011) considered knowledge elements as the basic element of knowledge
organization and constructed a knowledge organization neural system based on knowledge
element linkage. Wen Tingxiao et al. (2007) analyzed the value and difficulties in the Chinese text
knowledge element construction. They pointed out that the Chinese word segmentation would be
the bottleneck in the area. Word segmentation is an elementary process in Chinese text information
processing. However, it is not the key difficulty in knowledge element extraction. In the English
knowledge element extraction, the same problems remain. In determining knowledge elements,
we need to combine several words to form a knowledge element concept. For example, knowledge
management is a knowledge element unit while only or management alone cannot represent any
meaningful knowledge element.