The ontology merging algorithm based on concept
similarity can be divided into five steps:
(1) Concept processing: firstly, describe the two source
ontology to be merged in the unified language with
protégé-2000 to eliminate differences between the two
ontology; then, parse the different ontology separately with
Jean analyzing tool, and get the corresponding formal
context with the parse results.
(2) Analyze the formal context and concepts to be
merged according to the WordNet ontology, and get the
semantic relations of concepts. Then calculate the concept
similarity with equation (6), and get the relations between
concept and attribute, and the corresponding formal context.
(3) According to the formal context in step (2), the
concept lattice attribute reduction algorithm based on
mapping [10] is applied to delete the unnecessary attribute
[11]. In this way, it reduces the concept lattice to reflect the
nature of concept in the simplest structure.
(4) Construct the concept lattice with the formal context
reduced in step (3).
(5) Construct the ontology with the concept lattice
constructed in step (4).