Phase 1: Knowledge Extraction/Reuse

Phase 1: Knowledge Extraction/Reuse. As Figure 2 shows,
this phase is data gathering and pre-processing by adopting
abstraction techniques, designed abstraction algorithms and
mapping rules [17], and reusing knowledge bases. Specifically,
the extraction part works as following description. Firstly, it
determines the objective of task and selects relevant documents
as raw data. Then the domain vocabulary is extracted from the
text data supported by abstraction algorithms. Last, the
extracted domain vocabulary is mapped into the ontology
format to be the domain knowledge/information according to
designed mapping rules. Because building an ontology from
scratch is not only time consuming but also limited to gathered
resources, moreover, the ontology based domain knowledge is
reusable, thus, it is more efficient to reuse existing domain
ontologies to assist the construction of specific domain
knowledge base. In particular, there are two circumstances in
the knowledge reuse: (1) if a knowledge ontology exists for the
required domain but is not up to date, it requires a smaller scale
knowledge extraction to get the latest information and then
merges the extracted information into the existing domain
ontology to form the requisite knowledge base; and (2) if there
is a knowledge ontology extracted recently for the required
domain, the existing domain ontology will be reused directly as
the knowledge base for the subsequent idea generation.